victoriametrics.com

vmctl

VictoriaMetrics command-line tool (vmctl) provides the following list of actions:

To see the full list of supported actions run the following command:

$ ./vmctl --help
NAME:
   vmctl - VictoriaMetrics command-line tool

USAGE:
   vmctl [global options] command [command options] [arguments...]

COMMANDS:
   opentsdb    Migrate timeseries from OpenTSDB
   influx      Migrate timeseries from InfluxDB
   prometheus  Migrate timeseries from Prometheus
   vm-native   Migrate time series between VictoriaMetrics installations via native binary format
   remote-read Migrate timeseries by Prometheus remote read protocol
   verify-block  Verifies correctness of data blocks exported via VictoriaMetrics Native format. See https://docs.victoriametrics.com/#how-to-export-data-in-native-format

Each command has its own unique set of flags specific (e.g. prefixed with influx- for influx) to the data source and common list of flags for destination (prefixed with vm- for VictoriaMetrics):

$ ./vmctl influx --help
OPTIONS:
   --influx-addr value              InfluxDB server addr (default: "http://localhost:8086")
   --influx-user value              InfluxDB user [$INFLUX_USERNAME]
...
   --vm-addr vmctl                             VictoriaMetrics address to perform import requests.
Should be the same as --httpListenAddr value for single-node version or vminsert component.
When importing into the clustered version do not forget to set additionally --vm-account-id flag.
Please note, that vmctl performs initial readiness check for the given address by checking `/health` endpoint. (default: "http://localhost:8428")
   --vm-user value        VictoriaMetrics username for basic auth [$VM_USERNAME]
   --vm-password value    VictoriaMetrics password for basic auth [$VM_PASSWORD]

When doing a migration user needs to specify flags for source (where and how to fetch data) and for destination (where to migrate data). Every command has additional details and nuances, please see them below in corresponding sections.

For the destination flags see the full description by running the following command:

$ ./vmctl influx --help | grep vm-

Some flags like –vm-extra-label or –vm-significant-figures has additional sections with description below. Details about tweaking and adjusting settings are explained in Tuning section.

Please note, that if you’re going to import data into VictoriaMetrics cluster do not forget to specify the --vm-account-id flag. See more details for cluster version here.

Articles#

Migrating data from OpenTSDB#

vmctl supports the opentsdb mode to migrate data from OpenTSDB to VictoriaMetrics time-series database.

See ./vmctl opentsdb --help for details and full list of flags.

Important: OpenTSDB migration is not possible without a functioning meta table to search for metrics/series. Check in OpenTSDB config that appropriate options are activated and HBase meta tables are present. W/o them migration won’t work.

OpenTSDB migration works like so:

  1. Find metrics based on selected filters (or the default filter set ['a','b','c','d','e','f','g','h','i','j','k','l','m','n','o','p','q','r','s','t','u','v','w','x','y','z']):

    curl -Ss "http://opentsdb:4242/api/suggest?type=metrics&q=sys"

  2. Find series associated with each returned metric:

    curl -Ss "http://opentsdb:4242/api/search/lookup?m=system.load5&limit=1000000"

    Here results return field should not be empty. Otherwise, it means that meta tables are absent and needs to be turned on previously.

  3. Download data for each series in chunks defined in the CLI switches:

    -retention=sum-1m-avg:1h:90d means:

    • curl -Ss "http://opentsdb:4242/api/query?start=1h-ago&end=now&m=sum:1m-avg-none:system.load5\{host=host1\}"
    • curl -Ss "http://opentsdb:4242/api/query?start=2h-ago&end=1h-ago&m=sum:1m-avg-none:system.load5\{host=host1\}"
    • curl -Ss "http://opentsdb:4242/api/query?start=3h-ago&end=2h-ago&m=sum:1m-avg-none:system.load5\{host=host1\}"
    • …
    • curl -Ss "http://opentsdb:4242/api/query?start=2160h-ago&end=2159h-ago&m=sum:1m-avg-none:system.load5\{host=host1\}"

This means that we must stream data from OpenTSDB to VictoriaMetrics in chunks. This is where concurrency for OpenTSDB comes in. We can query multiple chunks at once, but we shouldn’t perform too many chunks at a time to avoid overloading the OpenTSDB cluster.

$ ./vmctl opentsdb --otsdb-addr http://opentsdb:4242/ --otsdb-retentions sum-1m-avg:1h:1d --otsdb-filters system --otsdb-normalize --vm-addr http://victoria:8428/
OpenTSDB import mode
2021/04/09 11:52:50 Will collect data starting at TS 1617990770
2021/04/09 11:52:50 Loading all metrics from OpenTSDB for filters:  [system]
Found 9 metrics to import. Continue? [Y/n]
2021/04/09 11:52:51 Starting work on system.load1
23 / 402200 [>____________________________________________________________________________________________] 0.01% 2 p/s

Where :8428 is Prometheus port of VictoriaMetrics.

For clustered VictoriaMetrics setup --vm-account-id flag needs to be added, for example:

$ ./vmctl opentsdb --otsdb-addr http://opentsdb:4242/ --otsdb-retentions sum-1m-avg:1h:1d --otsdb-filters system --otsdb-normalize --vm-addr http://victoria:8480/ --vm-account-id 0

This time :8480 port is vminsert/Prometheus input port.

Retention strings#

Starting with a relatively simple retention string (sum-1m-avg:1h:30d), let’s describe how this is converted into actual queries.

There are two essential parts of a retention string:

  1. aggregation
  2. windows/time ranges

Aggregation#

Retention strings essentially define the two levels of aggregation for our collected series.

sum-1m-avg would become:

  • First order: sum
  • Second order: 1m-avg-none
First Order Aggregations#

First-order aggregation addresses how to aggregate any un-mentioned tags.

This is, conceptually, directly opposite to how PromQL deals with tags. In OpenTSDB, if a tag isn’t explicitly mentioned, all values associated with that tag will be aggregated.

It is recommended to use sum for the first aggregation because it is relatively quick and should not cause any changes to the incoming data (because we collect each individual series).

Second Order Aggregations#

Second-order aggregation (1m-avg in our example) defines any windowing that should occur before returning the data

It is recommended to match the stat collection interval, so we again avoid transforming incoming data.

We do not allow for defining the “null value” portion of the rollup window (e.g. in the aggregation, 1m-avg-none, the user cannot change none), as the goal of this tool is to avoid modifying incoming data.

Windows#

There are two important windows we define in a retention string:

  1. the “chunk” range of each query
  2. The time range we will be querying on with that “chunk”

From our example, our windows are 1h:30d.

Window “chunks”#

The window 1h means that each individual query to OpenTSDB should only span 1 hour of time (e.g. start=2h-ago&end=1h-ago).

It is important to ensure this window somewhat matches the row size in HBase to help improve query times.

For example, if the query is hitting a rollup table with a 4-hour row size, we should set a chunk size of a multiple of 4 hours (e.g. 4h, 8h, etc.) to avoid requesting data across row boundaries. Landing on row boundaries allows for more consistent request times to HBase.

The default table created in HBase for OpenTSDB has a 1-hour row size, so if you aren’t sure on a correct row size to use, 1h is a reasonable choice.

Time range#

The time range 30d simply means we are asking for the last 30 days of data. This time range can be written using h, d, w, or y. (We can’t use m for month because it already means minute in time parsing).

Results of retention string#

The resultant queries that will be created, based on our example retention string of sum-1m-avg:1h:30d look like this:

http://opentsdb:4242/api/query?start=1h-ago&end=now&m=sum:1m-avg-none:<series>
http://opentsdb:4242/api/query?start=2h-ago&end=1h-ago&m=sum:1m-avg-none:<series>
http://opentsdb:4242/api/query?start=3h-ago&end=2h-ago&m=sum:1m-avg-none:<series>
...
http://opentsdb:4242/api/query?start=721h-ago&end=720h-ago&m=sum:1m-avg-none:<series>

Chunking the data like this means each individual query returns faster, so we can start populating data into VictoriaMetrics quicker.

Configuration#

Run the following command to get all configuration options:

./vmctl opentsdb --help

Restarting OpenTSDB migrations#

One important note for OpenTSDB migration: Queries/HBase scans can “get stuck” within OpenTSDB itself. This can cause instability and performance issues within an OpenTSDB cluster, so stopping the migrator to deal with it may be necessary. Because of this, we provide the timestamp we started collecting data from at the beginning of the run. You can stop and restart the importer using this “hard timestamp” to ensure you collect data from the same time range over multiple runs.

Migrating data from InfluxDB (1.x)#

vmctl supports the influx mode for migrating data from InfluxDB to VictoriaMetrics time-series database.

See ./vmctl influx --help for details and full list of flags.

To use migration tool please specify the InfluxDB address --influx-addr, the database --influx-database and VictoriaMetrics address --vm-addr. Flag --vm-addr for single-node VM is usually equal to --httpListenAddr, and for cluster version is equal to --httpListenAddr flag of vminsert component. Please note, that vmctl performs initial readiness check for the given address by checking /health endpoint. For cluster version it is additionally required to specify the --vm-account-id flag. See more details for cluster version here.

As soon as required flags are provided and all endpoints are accessible, vmctl will start the InfluxDB scheme exploration. Basically, it just fetches all fields and timeseries from the provided database and builds up registry of all available timeseries. Then vmctl sends fetch requests for each timeseries to InfluxDB one by one and pass results to VM importer. VM importer then accumulates received samples in batches and sends import requests to VM.

The importing process example for local installation of InfluxDB(http://localhost:8086) and single-node VictoriaMetrics(http://localhost:8428):

./vmctl influx --influx-database benchmark
InfluxDB import mode
2020/01/18 20:47:11 Exploring scheme for database "benchmark"
2020/01/18 20:47:11 fetching fields: command: "show field keys"; database: "benchmark"; retention: "autogen"
2020/01/18 20:47:11 found 10 fields
2020/01/18 20:47:11 fetching series: command: "show series "; database: "benchmark"; retention: "autogen"
Found 40000 timeseries to import. Continue? [Y/n] y
40000 / 40000 [----------------------------------------------------------------------------------------] 100.00% 21 p/s
2020/01/18 21:19:00 Import finished!
2020/01/18 21:19:00 VictoriaMetrics importer stats:
  idle duration: 13m51.461434876s;
  time spent while importing: 17m56.923899847s;
  total samples: 345600000;
  samples/s: 320914.04;
  total bytes: 5.9 GB;
  bytes/s: 5.4 MB;
  import requests: 40001;
2020/01/18 21:19:00 Total time: 31m48.467044016s

Data mapping#

Vmctl maps InfluxDB data the same way as VictoriaMetrics does by using the following rules:

  • influx-database arg is mapped into db label value unless db tag exists in the InfluxDB line. If you want to skip this mapping just enable flag influx-skip-database-label.
  • Field names are mapped to time series names prefixed with {measurement}{separator} value, where {separator} equals to _ by default. It can be changed with --influx-measurement-field-separator command-line flag.
  • Field values are mapped to time series values.
  • Tags are mapped to Prometheus labels format as-is.

For example, the following InfluxDB line:

foo,tag1=value1,tag2=value2 field1=12,field2=40

is converted into the following Prometheus format data points:

foo_field1{tag1="value1", tag2="value2"} 12
foo_field2{tag1="value1", tag2="value2"} 40

Configuration#

Run the following command to get all configuration options:

./vmctl influx --help

Filtering#

The filtering consists of two parts: timeseries and time. The first step of application is to select all available timeseries for given database and retention. User may specify additional filtering condition via --influx-filter-series flag. For example:

./vmctl influx --influx-database benchmark \
  --influx-filter-series "on benchmark from cpu where hostname='host_1703'"
InfluxDB import mode
2020/01/26 14:23:29 Exploring scheme for database "benchmark"
2020/01/26 14:23:29 fetching fields: command: "show field keys"; database: "benchmark"; retention: "autogen"
2020/01/26 14:23:29 found 12 fields
2020/01/26 14:23:29 fetching series: command: "show series on benchmark from cpu where hostname='host_1703'"; database: "benchmark"; retention: "autogen"
Found 10 timeseries to import. Continue? [Y/n]

The timeseries select query would be following: fetching series: command: "show series on benchmark from cpu where hostname='host_1703'"; database: "benchmark"; retention: "autogen"

The second step of filtering is a time filter and it applies when fetching the datapoints from Influx. Time filtering may be configured with two flags:

  • –influx-filter-time-start
  • –influx-filter-time-end Here’s an example of importing timeseries for one day only: ./vmctl influx --influx-database benchmark --influx-filter-series "where hostname='host_1703'" --influx-filter-time-start "2020-01-01T10:07:00Z" --influx-filter-time-end "2020-01-01T15:07:00Z"

Please see more about time filtering here.

Migrating data from InfluxDB (2.x)#

Migrating data from InfluxDB v2.x is not supported yet (#32). You may find useful a 3rd party solution for this - https://github.com/jonppe/influx_to_victoriametrics.

Migrating data from Promscale#

Promscale supports Prometheus Remote Read API. To migrate historical data from Promscale to VictoriaMetrics we recommend using vmctl in remote-read mode.

See the example of migration command below:

./vmctl remote-read --remote-read-src-addr=http://<promscale>:9201/read \
                    --remote-read-step-interval=day \
                    --remote-read-use-stream=false \ # promscale doesn't support streaming
                    --vm-addr=http://<victoriametrics>:8428 \
                    --remote-read-filter-time-start=2023-08-21T00:00:00Z \
                    --remote-read-disable-path-append=true # promscale has custom remote read API HTTP path
Selected time range "2023-08-21 00:00:00 +0000 UTC" - "2023-08-21 14:11:41.561979 +0000 UTC" will be split into 1 ranges according to "day" step. Continue? [Y/n] y
VM worker 0:↙ 82831 samples/s                                                                                                                                                                        
VM worker 1:↙ 54378 samples/s                                                                                                                                                                        
VM worker 2:↙ 121616 samples/s                                                                                                                                                                       
VM worker 3:↙ 59164 samples/s                                                                                                                                                                        
VM worker 4:↙ 59220 samples/s                                                                                                                                                                        
VM worker 5:↙ 102072 samples/s                                                                                                                                                                       
Processing ranges: 1 / 1 [████████████████████████████████████████████████████████████████████████████████████] 100.00%
2023/08/21 16:11:55 Import finished!
2023/08/21 16:11:55 VictoriaMetrics importer stats:
  idle duration: 0s;
  time spent while importing: 14.047045459s;
  total samples: 262111;
  samples/s: 18659.51;
  total bytes: 5.3 MB;
  bytes/s: 376.4 kB;
  import requests: 6;
  import requests retries: 0;
2023/08/21 16:11:55 Total time: 14.063458792s

Here we specify the full path to Promscale’s Remote Read API via --remote-read-src-addr, and disable auto-path appending via --remote-read-disable-path-append cmd-line flags. This is necessary, as Promscale has a different to Prometheus API path. Promscale doesn’t support stream mode for Remote Read API, so we disable it via --remote-read-use-stream=false.

Migrating data from Prometheus#

vmctl supports the prometheus mode for migrating data from Prometheus to VictoriaMetrics time-series database. Migration is based on reading Prometheus snapshot, which is basically a hard-link to Prometheus data files.

See ./vmctl prometheus --help for details and full list of flags. Also see Prometheus related articles here.

To use migration tool please specify the file path to Prometheus snapshot --prom-snapshot (see how to make a snapshot here) and VictoriaMetrics address --vm-addr. Please note, that vmctl do not make a snapshot from Prometheus, it uses an already prepared snapshot. More about Prometheus snapshots may be found here and here. Flag --vm-addr for single-node VM is usually equal to --httpListenAddr, and for cluster version is equal to --httpListenAddr flag of vminsert component. Please note, that vmctl performs initial readiness check for the given address by checking /health endpoint. For cluster version it is additionally required to specify the --vm-account-id flag. See more details for cluster version here.

As soon as required flags are provided and all endpoints are accessible, vmctl will start the Prometheus snapshot exploration. Basically, it just fetches all available blocks in provided snapshot and read the metadata. It also does initial filtering by time if flags --prom-filter-time-start or --prom-filter-time-end were set. The exploration procedure prints some stats from read blocks. Please note that stats are not taking into account timeseries or samples filtering. This will be done during importing process.

The importing process takes the snapshot blocks revealed from Explore procedure and processes them one by one accumulating timeseries and samples. Please note, that vmctl relies on responses from InfluxDB on this stage, so ensure that Explore queries are executed without errors or limits. Please see this issue for details. The data processed in chunks and then sent to VM.

The importing process example for local installation of Prometheus and single-node VictoriaMetrics(http://localhost:8428):

./vmctl prometheus --prom-snapshot=/path/to/snapshot \
  --vm-concurrency=1 \
  --vm-batch-size=200000 \
  --prom-concurrency=3
Prometheus import mode
Prometheus snapshot stats:
  blocks found: 14;
  blocks skipped: 0;
  min time: 1581288163058 (2020-02-09T22:42:43Z);
  max time: 1582409128139 (2020-02-22T22:05:28Z);
  samples: 32549106;
  series: 27289.
Found 14 blocks to import. Continue? [Y/n] y
14 / 14 [-------------------------------------------------------------------------------------------] 100.00% 0 p/s
2020/02/23 15:50:03 Import finished!
2020/02/23 15:50:03 VictoriaMetrics importer stats:
  idle duration: 6.152953029s;
  time spent while importing: 44.908522491s;
  total samples: 32549106;
  samples/s: 724786.84;
  total bytes: 669.1 MB;
  bytes/s: 14.9 MB;
  import requests: 323;
  import requests retries: 0;
2020/02/23 15:50:03 Total time: 51.077451066s

Data mapping#

VictoriaMetrics has very similar data model to Prometheus and supports RemoteWrite integration. So no data changes will be applied.

Configuration#

Run the following command to get all configuration options:

./vmctl prometheus --help

Filtering#

The filtering consists of three parts: by timeseries and time.

Filtering by time may be configured via flags --prom-filter-time-start and --prom-filter-time-end in RFC3339 format. This filter applied twice: to drop blocks out of range and to filter timeseries in blocks with overlapping time range.

Example of applying time filter:

./vmctl prometheus --prom-snapshot=/path/to/snapshot \
  --prom-filter-time-start=2020-02-07T00:07:01Z \
  --prom-filter-time-end=2020-02-11T00:07:01Z
Prometheus import mode
Prometheus snapshot stats:
  blocks found: 2;
  blocks skipped: 12;
  min time: 1581288163058 (2020-02-09T22:42:43Z);
  max time: 1581328800000 (2020-02-10T10:00:00Z);
  samples: 1657698;
  series: 3930.
Found 2 blocks to import. Continue? [Y/n] y

Please notice, that total amount of blocks in provided snapshot is 14, but only 2 of them were in provided time range. So other 12 blocks were marked as skipped. The amount of samples and series is not taken into account, since this is heavy operation and will be done during import process.

Filtering by timeseries is configured with following flags:

  • --prom-filter-label - the label name, e.g. __name__ or instance;
  • --prom-filter-label-value - the regular expression to filter the label value. By default, matches all .*

For example:

./vmctl prometheus --prom-snapshot=/path/to/snapshot \
  --prom-filter-label="__name__" \
  --prom-filter-label-value="promhttp.*" \
  --prom-filter-time-start=2020-02-07T00:07:01Z \
  --prom-filter-time-end=2020-02-11T00:07:01Z
Prometheus import mode
Prometheus snapshot stats:
  blocks found: 2;
  blocks skipped: 12;
  min time: 1581288163058 (2020-02-09T22:42:43Z);
  max time: 1581328800000 (2020-02-10T10:00:00Z);
  samples: 1657698;
  series: 3930.
Found 2 blocks to import. Continue? [Y/n] y
14 / 14 [-----------------------------------------------------------------------------------------------] 100.00% ? p/s
2020/02/23 15:51:07 Import finished!
2020/02/23 15:51:07 VictoriaMetrics importer stats:
  idle duration: 0s;
  time spent while importing: 37.415461ms;
  total samples: 10128;
  samples/s: 270690.24;
  total bytes: 195.2 kB;
  bytes/s: 5.2 MB;
  import requests: 2;
  import requests retries: 0;
2020/02/23 15:51:07 Total time: 7.153158218s

Migrating data by remote read protocol#

vmctl provides the remote-read mode for migrating data from remote databases supporting Prometheus remote read API. Remote read API has two implementations of remote read API: default (SAMPLES) and streamed (STREAMED_XOR_CHUNKS). Streamed version is more efficient but has lower adoption (e.g. Promscale doesn’t support it).

See ./vmctl remote-read --help for details and the full list of flags.

To start the migration process configure the following flags:

  1. --remote-read-src-addr - data source address to read from;
  2. --vm-addr - VictoriaMetrics address to write to. For single-node VM is usually equal to --httpListenAddr, and for cluster version is equal to --httpListenAddr flag of vminsert component (for example http://<vminsert>:8480/insert/<accountID>/prometheus);
  3. --remote-read-filter-time-start - the time filter in RFC3339 format to select time series with timestamp equal or higher than provided value. E.g. ‘2020-01-01T20:07:00Z’;
  4. --remote-read-filter-time-end - the time filter in RFC3339 format to select time series with timestamp equal or smaller than provided value. E.g. ‘2020-01-01T20:07:00Z’. Current time is used when omitted.;
  5. --remote-read-step-interval - split export data into chunks. Valid values are month, day, hour, minute;
  6. --remote-read-use-stream - defines whether to use SAMPLES or STREAMED_XOR_CHUNKS mode. By default, is uses SAMPLES mode.

The importing process example for local installation of Prometheus and single-node VictoriaMetrics(http://localhost:8428):

./vmctl remote-read \
--remote-read-src-addr=http://<prometheus>:9091 \
--remote-read-filter-time-start=2021-10-18T00:00:00Z \
--remote-read-step-interval=hour \
--vm-addr=http://<victoria-metrics>:8428 \
--vm-concurrency=6

Split defined times into 8798 ranges to import. Continue? [Y/n]
VM worker 0:↘ 127177 samples/s
VM worker 1:↘ 140137 samples/s
VM worker 2:↘ 151606 samples/s
VM worker 3:↘ 130765 samples/s
VM worker 4:↘ 131904 samples/s
VM worker 5:↘ 132693 samples/s
Processing ranges: 8798 / 8798 [██████████████████████████████████████████████████████████████████████████████] 100.00%
2022/10/19 16:45:37 Import finished!
2022/10/19 16:45:37 VictoriaMetrics importer stats:
  idle duration: 6m57.793987511s;
  time spent while importing: 1m18.463744801s;
  total samples: 25348208;
  samples/s: 323056.31;
  total bytes: 669.7 MB;
  bytes/s: 8.5 MB;
  import requests: 127;
  import requests retries: 0;
2022/10/19 16:45:37 Total time: 1m19.406283424s

Migrating big volumes of data may result in remote read client reaching the timeout. Consider increasing the value of --remote-read-http-timeout (default 5m) command-line flag when seeing timeouts or context canceled errors.

Filtering#

The filtering consists of two parts: by labels and time.

Filtering by time can be configured via flags --remote-read-filter-time-start and --remote-read-filter-time-end in RFC3339 format.

Filtering by labels can be configured via flags --remote-read-filter-label and --remote-read-filter-label-value. For example, --remote-read-filter-label=tenant and --remote-read-filter-label-value="team-eu" will select only series with tenant="team-eu" label-value pair.

Migrating data from Thanos#

Thanos uses the same storage engine as Prometheus and the data layout on-disk should be the same. That means vmctl in mode prometheus may be used for Thanos historical data migration as well. These instructions may vary based on the details of your Thanos configuration. Please read carefully and verify as you go. We assume you’re using Thanos Sidecar on your Prometheus pods, and that you have a separate Thanos Store installation.

Current data#

  1. For now, keep your Thanos Sidecar and Thanos-related Prometheus configuration, but add this to also stream metrics to VictoriaMetrics:

    remote_write:
- url: http://victoria-metrics:8428/api/v1/write

  2. Make sure VM is running, of course. Now check the logs to make sure that Prometheus is sending and VM is receiving. In Prometheus, make sure there are no errors. On the VM side, you should see messages like this:

    2020-04-27T18:38:46.474Z info VictoriaMetrics/lib/storage/partition.go:207 creating a partition "2020_04" with smallPartsPath="/victoria-metrics-data/data/small/2020_04", bigPartsPath="/victoria-metrics-data/data/big/2020_04"
2020-04-27T18:38:46.506Z info VictoriaMetrics/lib/storage/partition.go:222 partition "2020_04" has been created

  3. Now just wait. Within two hours, Prometheus should finish its current data file and hand it off to Thanos Store for long term storage.

Historical data#

Let’s assume your data is stored on S3 served by minio. You first need to copy that out to a local filesystem, then import it into VM using vmctl in prometheus mode.

  1. Copy data from minio.

    1. Run the minio/mc Docker container.
    2. mc config host add minio http://minio:9000 accessKey secretKey, substituting appropriate values for the last 3 items.
    3. mc cp -r minio/prometheus thanos-data

  2. Import using vmctl.

    1. Follow the instructions to compile vmctl on your machine.
    2. Use prometheus mode to import data:
    vmctl prometheus --prom-snapshot thanos-data --vm-addr http://victoria-metrics:8428

Remote read protocol#

Currently, Thanos doesn’t support streaming remote read protocol. It is recommended to use thanos-remote-read a proxy, that allows exposing any Thanos service (or anything that exposes gRPC StoreAPI e.g. Querier) via Prometheus remote read protocol.

If you want to migrate data, you should run thanos-remote-read proxy and define the Thanos store address ./thanos-remote-read -store 127.0.0.1:19194. It is important to know that store flag is Thanos Store API gRPC endpoint. Also, it is important to know that thanos-remote-read proxy doesn’t support stream mode. When you run thanos-remote-read proxy, it exposes port to serve HTTP on 10080 by default.

The importing process example for local installation of Thanos and single-node VictoriaMetrics(http://localhost:8428):

./vmctl remote-read \
--remote-read-src-addr=http://127.0.0.1:10080 \
--remote-read-filter-time-start=2021-10-18T00:00:00Z \
--remote-read-step-interval=hour \
--vm-addr=http://127.0.0.1:8428 \
--vm-concurrency=6

On the thanos-remote-read proxy side you will see logs like:

ts=2022-10-19T15:05:04.193916Z caller=main.go:278 level=info traceID=00000000000000000000000000000000 msg="thanos request" request="min_time:1666180800000 max_time:1666184399999 matchers:<type:RE value:\".*\" > aggregates:RAW "
ts=2022-10-19T15:05:04.468852Z caller=main.go:278 level=info traceID=00000000000000000000000000000000 msg="thanos request" request="min_time:1666184400000 max_time:1666187999999 matchers:<type:RE value:\".*\" > aggregates:RAW "
ts=2022-10-19T15:05:04.553914Z caller=main.go:278 level=info traceID=00000000000000000000000000000000 msg="thanos request" request="min_time:1666188000000 max_time:1666191364863 matchers:<type:RE value:\".*\" > aggregates:RAW "

And when process will finish you will see:

Split defined times into 8799 ranges to import. Continue? [Y/n]
VM worker 0:↓ 98183 samples/s
VM worker 1:↓ 114640 samples/s
VM worker 2:↓ 131710 samples/s
VM worker 3:↓ 114256 samples/s
VM worker 4:↓ 105671 samples/s
VM worker 5:↓ 124000 samples/s
Processing ranges: 8799 / 8799 [██████████████████████████████████████████████████████████████████████████████] 100.00%
2022/10/19 18:05:07 Import finished!
2022/10/19 18:05:07 VictoriaMetrics importer stats:
  idle duration: 52m13.987637229s;
  time spent while importing: 9m1.728983776s;
  total samples: 70836111;
  samples/s: 130759.32;
  total bytes: 2.2 GB;
  bytes/s: 4.0 MB;
  import requests: 356;
  import requests retries: 0;
2022/10/19 18:05:07 Total time: 9m2.607521618s

Migrating data from Cortex#

Cortex has an implementation of the Prometheus remote read protocol. That means vmctl in mode remote-read may also be used for Cortex historical data migration. These instructions may vary based on the details of your Cortex configuration. Please read carefully and verify as you go.

Remote read protocol#

If you want to migrate data, you should check your cortex configuration in the section

api:
  prometheus_http_prefix:

If you defined some prometheus prefix, you should use it when you define flag --remote-read-src-addr=http://127.0.0.1:9009/{prometheus_http_prefix}. By default, Cortex uses the prometheus path prefix, so you should define the flag --remote-read-src-addr=http://127.0.0.1:9009/prometheus.

It is important to know that Cortex doesn’t support the stream mode. When you run Cortex, it exposes a port to serve HTTP on 9009 by default.

The importing process example for the local installation of Cortex and single-node VictoriaMetrics(http://localhost:8428):

./vmctl remote-read \ 
--remote-read-src-addr=http://127.0.0.1:9009/prometheus \
--remote-read-filter-time-start=2021-10-18T00:00:00Z \
--remote-read-step-interval=hour \
--vm-addr=http://127.0.0.1:8428 \
--vm-concurrency=6

And when the process finishes, you will see the following:

Split defined times into 8842 ranges to import. Continue? [Y/n]
VM worker 0:↗ 3863 samples/s
VM worker 1:↗ 2686 samples/s
VM worker 2:↗ 2620 samples/s
VM worker 3:↗ 2705 samples/s
VM worker 4:↗ 2643 samples/s
VM worker 5:↗ 2593 samples/s
Processing ranges: 8842 / 8842 [█████████████████████████████████████████████████████████████████████████████] 100.00%
2022/10/21 12:09:49 Import finished!
2022/10/21 12:09:49 VictoriaMetrics importer stats:
  idle duration: 0s;
  time spent while importing: 3.82640757s;
  total samples: 160232;
  samples/s: 41875.31;
  total bytes: 11.3 MB;
  bytes/s: 3.0 MB;
  import requests: 6;
  import requests retries: 0;
2022/10/21 12:09:49 Total time: 4.71824253s

It is important to know that if you run your Cortex installation in multi-tenant mode, remote read protocol requires an Authentication header like X-Scope-OrgID. You can define it via the flag --remote-read-headers=X-Scope-OrgID:demo

Migrating data from Mimir#

Mimir has similar implementation as Cortex and supports Prometheus remote read API. That means historical data from Mimir can be migrated via vmctl in mode remote-read mode. The instructions for data migration via vmctl vary based on the details of your Mimir configuration. Please read carefully and verify as you go.

Remote read protocol#

By default, Mimir uses the prometheus path prefix so specifying the source should be as simple as --remote-read-src-addr=http://<mimir>:9009/prometheus. But if prefix was overriden via prometheus_http_prefix, then source address should be updated to --remote-read-src-addr=http://<mimir>:9009/{prometheus_http_prefix}.

Mimir supports streamed remote read API, so it is recommended setting --remote-read-use-stream=true flag for better performance and resource usage.

When you run Mimir, it exposes a port to serve HTTP on 8080 by default.

Next example of the local installation was in multi-tenant mode (3 instances of Mimir) with nginx as load balancer. Load balancer expose single port :9090.

As you can see in the example we call :9009 instead of :8080 because of proxy.

The importing process example for the local installation of Mimir and single-node VictoriaMetrics(http://localhost:8428):

./vmctl remote-read 
--remote-read-src-addr=http://<mimir>:9009/prometheus \
--remote-read-filter-time-start=2021-10-18T00:00:00Z \
--remote-read-step-interval=hour \
--remote-read-headers=X-Scope-OrgID:demo \
--remote-read-use-stream=true \
--vm-addr=http://<victoria-metrics>:8428 \

And when the process finishes, you will see the following:

Split defined times into 8847 ranges to import. Continue? [Y/n]
VM worker 0:→ 12176 samples/s
VM worker 1:→ 11918 samples/s
VM worker 2:→ 11261 samples/s
VM worker 3:→ 12861 samples/s
VM worker 4:→ 11096 samples/s
VM worker 5:→ 11575 samples/s
Processing ranges: 8847 / 8847 [█████████████████████████████████████████████████████████████████████████████] 100.00%
2022/10/21 17:22:23 Import finished!
2022/10/21 17:22:23 VictoriaMetrics importer stats:
  idle duration: 0s;
  time spent while importing: 15.379614356s;
  total samples: 81243;
  samples/s: 5282.51;
  total bytes: 6.1 MB;
  bytes/s: 397.8 kB;
  import requests: 6;
  import requests retries: 0;
2022/10/21 17:22:23 Total time: 16.287405248s

It is important to know that if you run your Mimir installation in multi-tenant mode, remote read protocol requires an Authentication header like X-Scope-OrgID. You can define it via the flag --remote-read-headers=X-Scope-OrgID:demo

Migrating data from VictoriaMetrics#

The simplest way to migrate data between VM instances is to copy data between instances.

vmctl uses native binary protocol (available since 1.42.0 release) to migrate data between VM instances: single to single, cluster to cluster, single to cluster and vice versa.

See ./vmctl vm-native --help for details and full list of flags.

Migration in vm-native mode takes two steps:

  1. Explore the list of the metrics to migrate via api/v1/label/__name__/values API;
  2. Migrate explored metrics one-by-one with specified --vm-concurrency.
./vmctl vm-native \
    --vm-native-src-addr=http://127.0.0.1:8481/select/0/prometheus \ # migrate from
    --vm-native-dst-addr=http://localhost:8428 \                     # migrate to
    --vm-native-filter-time-start='2022-11-20T00:00:00Z' \           # starting from
    --vm-native-filter-match='{__name__!~"vm_.*"}'                   # filter out metrics matching the selector
VictoriaMetrics Native import mode

2023/03/02 09:22:02 Initing import process from "http://127.0.0.1:8481/select/0/prometheus/api/v1/export/native" 
                    to "http://localhost:8428/api/v1/import/native" with filter 
        filter: match[]={__name__!~"vm_.*"}
        start: 2022-11-20T00:00:00Z
2023/03/02 09:22:02 Exploring metrics...
Found 9 metrics to import. Continue? [Y/n] 
2023/03/02 09:22:04 Requests to make: 9
Requests to make: 9 / 9 [█████████████████████████████████████████████████████████████████████████████] 100.00%
2023/03/02 09:22:06 Import finished!
2023/03/02 09:22:06 VictoriaMetrics importer stats:
  time spent while importing: 3.632638875s;
  total bytes: 7.8 MB;
  bytes/s: 2.1 MB;
  requests: 9;
  requests retries: 0;
2023/03/02 09:22:06 Total time: 3.633127625s

To disable explore phase and switch to the old way of data migration via single connection use --vm-native-disable-per-metric-migration cmd-line flag. Please note, in this mode vmctl won’t be able to retry failed requests.

Importing tips:

  1. vmctl acts as a proxy between src and dst. It doesn’t use much of CPU or RAM, but network connection between src=>vmctl=>dst should be as fast as possible for improving the migration speed.
  2. Migrating big volumes of data may result in reaching the safety limits on src side. Please verify that -search.maxExportDuration and -search.maxExportSeries were set with proper values for src. If hitting the limits, follow the recommendations here. If hitting the number of matching timeseries exceeds... error, adjust filters to match less time series or update -search.maxSeries command-line flag on vmselect/vmsingle;
  3. Using smaller intervals via --vm-native-step-interval cmd-line flag can reduce the number of matched series per-request for sources with high churn rate. See more about step interval here.
  4. Migrating all the metrics from one VM to another may collide with existing application metrics (prefixed with vm_) at destination and lead to confusion when using official Grafana dashboards. To avoid such situation try to filter out VM process metrics via --vm-native-filter-match='{__name__!~"vm_.*"}' flag.
  5. Migrating data with overlapping time range or via unstable network can produce duplicates series at destination. To avoid duplicates set -dedup.minScrapeInterval=1ms for vmselect/vmstorage at the destination. This will instruct vmselect/vmstorage to ignore duplicates with identical timestamps. Ignore this recommendation if you already have -dedup.minScrapeInterval set to 1ms or higher values at destination.
  6. When migrating data from one VM cluster to another, consider using cluster-to-cluster mode. Or manually specify addresses according to URL format:
    # Migrating from cluster specific tenantID to single
--vm-native-src-addr=http://<src-vmselect>:8481/select/0/prometheus
--vm-native-dst-addr=http://<dst-vmsingle>:8428

# Migrating from single to cluster specific tenantID
--vm-native-src-addr=http://<src-vmsingle>:8428
--vm-native-dst-addr=http://<dst-vminsert>:8480/insert/0/prometheus

# Migrating single to single
--vm-native-src-addr=http://<src-vmsingle>:8428
--vm-native-dst-addr=http://<dst-vmsingle>:8428

# Migrating cluster to cluster for specific tenant ID
--vm-native-src-addr=http://<src-vmselect>:8481/select/0/prometheus
--vm-native-dst-addr=http://<dst-vminsert>:8480/insert/0/prometheus
  7. Migrating data from VM cluster which had replication (-replicationFactor > 1) enabled won’t produce the same amount of data copies for the destination database, and will result only in creating duplicates. To remove duplicates, destination database need to be configured with -dedup.minScrapeInterval=1ms. To restore the replication factor the destination vminsert component need to be configured with the according -replicationFactor value. See more about replication here.
  8. Migration speed can be adjusted via --vm-concurrency cmd-line flag, which controls the number of concurrent workers busy with processing. Please note, that each worker can load up to a single vCPU core on VictoriaMetrics. So try to set it according to allocated CPU resources of your VictoriaMetrics destination installation.
  9. Migration is a backfilling process, so it is recommended to read Backfilling tips section.
  10. vmctl doesn’t provide relabeling or other types of labels management. Instead, use relabeling in VictoriaMetrics.
  11. vmctl supports --vm-native-src-headers and --vm-native-dst-headers to define headers sent with each request to the corresponding source address.
  12. vmctl supports --vm-native-disable-http-keep-alive to allow vmctl to use non-persistent HTTP connections to avoid error use of closed network connection when running a heavy export requests.

Using time-based chunking of migration#

It is possible to split the migration process into steps based on time via --vm-native-step-interval cmd-line flag. Supported values are: month, week, day, hour, minute. For example, when migrating 1 year of data with --vm-native-step-interval=month vmctl will execute it in 12 separate requests from the beginning of the interval to its end. To reverse the order set --vm-native-filter-time-reverse and migration will start from the newest to the oldest data. --vm-native-filter-time-start is required to be set when using --vm-native-step-interval.

It is recommended using default month step when migrating the data over the long time intervals. If you hit complexity limits on --vm-native-src-addr and can’t or don’t want to change them, try lowering the step interval to week, day or hour.

Usage example:

./vmctl vm-native \
    --vm-native-src-addr=http://127.0.0.1:8481/select/0/prometheus \ 
    --vm-native-dst-addr=http://localhost:8428 \
    --vm-native-filter-time-start='2022-11-20T00:00:00Z' \
    --vm-native-step-interval=month \
    --vm-native-filter-match='{__name__!~"vm_.*"}'    
VictoriaMetrics Native import mode

2023/03/02 09:18:05 Initing import process from "http://127.0.0.1:8481/select/0/prometheus/api/v1/export/native" to "http://localhost:8428/api/v1/import/native" with filter 
        filter: match[]={__name__!~"vm_.*"}
        start: 2022-11-20T00:00:00Z
2023/03/02 09:18:05 Exploring metrics...
Found 9 metrics to import. Continue? [Y/n] 
2023/03/02 09:18:07 Selected time range will be split into 5 ranges according to "month" step. Requests to make: 45.
Requests to make: 45 / 45 [██████████████████████████████████████████████████████████████████████████████████] 100.00%
2023/03/02 09:18:12 Import finished!
2023/03/02 09:18:12 VictoriaMetrics importer stats:
  time spent while importing: 7.111870667s;
  total bytes: 7.7 MB;
  bytes/s: 1.1 MB;
  requests: 45;
  requests retries: 0;
2023/03/02 09:18:12 Total time: 7.112405875s

Cluster-to-cluster migration mode#

Using cluster-to-cluster migration mode helps to migrate all tenants data in a single vmctl run.

Cluster-to-cluster uses /admin/tenants endpoint (available starting from v1.84.0) to discover list of tenants from source cluster.

To use this mode you need to set --vm-intercluster flag to true, --vm-native-src-addr flag to ‘http://vmselect:8481/’ and --vm-native-dst-addr value to http://vminsert:8480/:

  ./vmctl vm-native --vm-native-src-addr=http://127.0.0.1:8481/ \
  --vm-native-dst-addr=http://127.0.0.1:8480/ \
  --vm-native-filter-match='{__name__="vm_app_uptime_seconds"}' \
  --vm-native-filter-time-start='2023-02-01T00:00:00Z' \
  --vm-native-step-interval=day \  
  --vm-intercluster
  
VictoriaMetrics Native import mode
2023/02/28 10:41:42 Discovering tenants...
2023/02/28 10:41:42 The following tenants were discovered: [0:0 1:0 2:0 3:0 4:0]
2023/02/28 10:41:42 Initing import process from "http://127.0.0.1:8481/select/0:0/prometheus/api/v1/export/native" to "http://127.0.0.1:8480/insert/0:0/prometheus/api/v1/import/native" with filter 
        filter: match[]={__name__="vm_app_uptime_seconds"}
        start: 2023-02-01T00:00:00Z for tenant 0:0 
2023/02/28 10:41:42 Exploring metrics...
2023/02/28 10:41:42 Found 1 metrics to import 
2023/02/28 10:41:42 Selected time range will be split into 28 ranges according to "day" step. 
Requests to make for tenant 0:0: 28 / 28 [████████████████████████████████████████████████████████████████████] 100.00%

2023/02/28 10:41:45 Initing import process from "http://127.0.0.1:8481/select/1:0/prometheus/api/v1/export/native" to "http://127.0.0.1:8480/insert/1:0/prometheus/api/v1/import/native" with filter 
        filter: match[]={__name__="vm_app_uptime_seconds"}
        start: 2023-02-01T00:00:00Z for tenant 1:0 
2023/02/28 10:41:45 Exploring metrics...
2023/02/28 10:41:45 Found 1 metrics to import 
2023/02/28 10:41:45 Selected time range will be split into 28 ranges according to "day" step. Requests to make: 28 
Requests to make for tenant 1:0: 28 / 28 [████████████████████████████████████████████████████████████████████] 100.00%

...

2023/02/28 10:42:49 Import finished!
2023/02/28 10:42:49 VictoriaMetrics importer stats:
  time spent while importing: 1m6.714210417s;
  total bytes: 39.7 MB;
  bytes/s: 594.4 kB;
  requests: 140;
  requests retries: 0;
2023/02/28 10:42:49 Total time: 1m7.147971417s

Configuration#

Run the following command to get all configuration options:

./vmctl vm-native --help

Tuning#

Verifying exported blocks from VictoriaMetrics#

In this mode, vmctl allows verifying correctness and integrity of data exported via native format from VictoriaMetrics. You can verify exported data at disk before uploading it by vmctl verify-block command:

# export blocks from VictoriaMetrics
curl localhost:8428/api/v1/export/native -g -d 'match[]={__name__!=""}' -o exported_data_block
# verify block content
./vmctl verify-block exported_data_block
2022/03/30 18:04:50 verifying block at path="exported_data_block"
2022/03/30 18:04:50 successfully verified block at path="exported_data_block", blockCount=123786
2022/03/30 18:04:50 Total time: 100.108ms

Tuning#

InfluxDB mode#

The flag --influx-concurrency controls how many concurrent requests may be sent to InfluxDB while fetching timeseries. Please set it wisely to avoid InfluxDB overwhelming.

The flag --influx-chunk-size controls the max amount of datapoints to return in single chunk from fetch requests. Please see more details here. The chunk size is used to control InfluxDB memory usage, so it won’t OOM on processing large timeseries with billions of datapoints.

Prometheus mode#

The flag --prom-concurrency controls how many concurrent readers will be reading the blocks in snapshot. Since snapshots are just files on disk it would be hard to overwhelm the system. Please go with value equal to number of free CPU cores.

VictoriaMetrics importer#

The flag --vm-concurrency controls the number of concurrent workers that process the import requests to destination. Please note that each import request can load up to a single vCPU core on VictoriaMetrics. So try to set it according to allocated CPU resources of your VictoriaMetrics installation.

Importer stats#

After successful import vmctl prints some statistics for details. The important numbers to watch are following:

  • idle duration - shows time that importer spent while waiting for data from InfluxDB/Prometheus to fill up --vm-batch-size batch size. Value shows total duration across all workers configured via --vm-concurrency. High value may be a sign of too slow InfluxDB/Prometheus fetches or too high --vm-concurrency value. Try to improve it by increasing --<mode>-concurrency value or decreasing --vm-concurrency value.
  • import requests - shows how many import requests were issued to VM server. The import request is issued once the batch size(--vm-batch-size) is full and ready to be sent. Please prefer big batch sizes (50k-500k) to improve performance.
  • import requests retries - shows number of unsuccessful import requests. Non-zero value may be a sign of network issues or VM being overloaded. See the logs during import for error messages.

Silent mode#

By default vmctl waits confirmation from user before starting the import. If this is unwanted behavior and no user interaction required - pass -s flag to enable “silence” mode:

See below the example of vm-native migration process:

    -s Whether to run in silent mode. If set to true no confirmation prompts will appear. (default: false)

Significant figures#

vmctl allows to limit the number of significant figures before importing. For example, the average value for response size is 102.342305 bytes and it has 9 significant figures. If you ask a human to pronounce this value then with high probability value will be rounded to first 4 or 5 figures because the rest aren’t really that important to mention. In most cases, such a high precision is too much. Moreover, such values may be just a result of floating point arithmetic, create a false precision and result into bad compression ratio according to information theory.

vmctl provides the following flags for improving data compression:

  • --vm-round-digits flag for rounding processed values to the given number of decimal digits after the point. For example, --vm-round-digits=2 would round 1.2345 to 1.23. By default, the rounding is disabled.

  • --vm-significant-figures flag for limiting the number of significant figures in processed values. It takes no effect if set to 0 (by default), but set --vm-significant-figures=5 and 102.342305 will be rounded to 102.34.

The most common case for using these flags is to improve data compression for time series storing aggregation results such as average, rate, etc.

Adding extra labels#

vmctl allows to add extra labels to all imported series. It can be achieved with flag --vm-extra-label label=value. If multiple labels needs to be added, set flag for each label, for example, --vm-extra-label label1=value1 --vm-extra-label label2=value2. If timeseries already have label, that must be added with --vm-extra-label flag, flag has priority and will override label value from timeseries.

Rate limiting#

Limiting the rate of data transfer could help to reduce pressure on disk or on destination database. The rate limit may be set in bytes-per-second via --vm-rate-limit flag.

Please note, you can also use vmagent as a proxy between vmctl and destination with -remoteWrite.rateLimit flag enabled.

How to build#

It is recommended using binary releases - vmctl is located in vmutils-* archives there.

Development build#

  1. Install Go. The minimum supported version is Go 1.22.
  2. Run make vmctl from the root folder of the repository. It builds vmctl binary and puts it into the bin folder.

Production build#

  1. Install docker.
  2. Run make vmctl-prod from the root folder of the repository. It builds vmctl-prod binary and puts it into the bin folder.

Building docker images#

Run make package-vmctl. It builds victoriametrics/vmctl:<PKG_TAG> docker image locally. <PKG_TAG> is auto-generated image tag, which depends on source code in the repository. The <PKG_TAG> may be manually set via PKG_TAG=foobar make package-vmctl.

The base docker image is alpine but it is possible to use any other base image by setting it via <ROOT_IMAGE> environment variable. For example, the following command builds the image on top of scratch image:

ROOT_IMAGE=scratch make package-vmctl

ARM build#

ARM build may run on Raspberry Pi or on energy-efficient ARM servers.

Development ARM build#

  1. Install Go. The minimum supported version is Go 1.22.
  2. Run make vmctl-linux-arm or make vmctl-linux-arm64 from the root folder of the repository. It builds vmctl-linux-arm or vmctl-linux-arm64 binary respectively and puts it into the bin folder.

Production ARM build#

  1. Install docker.
  2. Run make vmctl-linux-arm-prod or make vmctl-linux-arm64-prod from the root folder of the repository. It builds vmctl-linux-arm-prod or vmctl-linux-arm64-prod binary respectively and puts it into the bin folder.

Hide

Back to Page