Commit graph

1605 commits

Author SHA1 Message Date
Aliaksandr Valialkin
eee6f1e56d lib/persistentqueue: make the persistent queue more durable against unclean shutdown (kill -9, OOM, hard reset)
The strategy is:

- Periodical flushing of inmemory blocks to files, so they aren't lost on unclean shutdown.
- Periodical syncing of metadata for persisted queues, so the metadata remains in sync with the persisted data.
- Automatic adjusting of too big chunk size when opening the queue. The chunk size may be bigger than the writer offset after unclean shutdown.
- Skipping of broken chunk file if it cannot be read.
- Fsyncing finalized chunk files.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/687
2020-09-16 18:13:24 +03:00
Aliaksandr Valialkin
6ce52e3702 lib/protoparser/vmimport: add more testcases for invalid timestamps and values
Updates https://github.com/VictoriaMetrics/vmctl/issues/25
2020-09-16 02:21:53 +03:00
Aliaksandr Valialkin
cd87ca303f lib/protoparser: report more errors for incorrect timestamps and/or values
Previously certain errors in timestamps and/or values could be silently skipped,
which could lead to samples with zero values stored in the database.

Updates https://github.com/VictoriaMetrics/vmctl/issues/25
2020-09-16 02:16:15 +03:00
Aliaksandr Valialkin
5c4e111b43 lib/protoparser/graphite: return error when value or timestamp cannot be properly parsed
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/99
2020-09-16 02:16:13 +03:00
Aliaksandr Valialkin
1587f83fa0 app/vmselect/netstorage: typo fix after 03dfccfbed 2020-09-16 00:10:33 +03:00
Aliaksandr Valialkin
2a86e2fb98 vendor: update github.com/valyala/fastjson from v1.5.4 to v1.5.5
This should properly parse `+Inf` values when importing JSON lines via `/api/v1/import`

Updates https://github.com/VictoriaMetrics/vmctl/issues/25
2020-09-16 00:09:05 +03:00
Aliaksandr Valialkin
d39a985d6b docs/Single-server-VictoriaMetrics.md: mention that VictoriaMetrics ignores NaN and Inf values during data ingestion
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/752
2020-09-15 23:41:10 +03:00
Aliaksandr Valialkin
03dfccfbed app/vmselect/netstorage: reduce memory usage when the time range from query touches big number of samples per each time series 2020-09-15 21:08:09 +03:00
Aliaksandr Valialkin
27cd5555e6 app/vmselect/netstorage: mention that RunParallel or Cancel must be called on the returned results from ProcessSearchQuery 2020-09-15 20:39:43 +03:00
Aliaksandr Valialkin
1f0c0b0f6b app/vmselect: typo fix in -search.maxStalenessInterval description 2020-09-15 14:25:34 +03:00
Aliaksandr Valialkin
0c1c1b79ba lib/promscrape: add a link to troubleshooting docs to error message when duplicate scrape target with identical labels is skipped 2020-09-15 14:16:20 +03:00
Aliaksandr Valialkin
5f04b00b2d docs/Articles.md: add a link to https://medium.com/miro-engineering/prometheus-high-availability-and-fault-tolerance-strategy-long-term-storage-with-victoriametrics-82f6f3f0409e 2020-09-15 12:29:44 +03:00
Aliaksandr Valialkin
ca08161b54 lib/promscrape: typo fix 2020-09-12 00:14:15 +03:00
Roman Khavronenko
e2b31590e6 vmalert: add Group name as label to generated alerts and timeseries (#761)
Solves #611
2020-09-11 23:41:12 +03:00
Roman Khavronenko
16e0bb496e vmalert: update groups on config reload only if changes detected (#759)
On config reload event `vmalert` reloads configuration for every group. While
it works for simple configurations, the more complex and heavy installations may
suffer from frequent config reloads.
The change introduces the `checksum` field for every group and is set to md5 hash
of yaml configuration. The checksum will change if on any change to group
definition like rules order or annotation change. Comparing the `checksum` field
on config reload event helps to detect if group should be updated.
The groups update is now done concurrently, so reload duration will be limited by
the slowest group now.

Partially solves #691 by improving config reload speed.
2020-09-11 23:41:12 +03:00
Aliaksandr Valialkin
e53235ac5c lib/promscrape: do not reset the remaining rows when pushing a part of data to remote storage during big scrapes
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/753

Thanks to @PerGon and @clmssz for help with debugging.
2020-09-11 23:38:17 +03:00
Aliaksandr Valialkin
b776a93608 app/vmselect/promql: support composite durations like Prometheus 2.21 does
The following durations are supported now: `1h5m35s` or `1s543ms`

See https://github.com/prometheus/prometheus/releases/tag/v2.21.0
and https://github.com/prometheus/prometheus/pull/7713
2020-09-11 22:17:24 +03:00
Aliaksandr Valialkin
114cf24b43 lib/promscrape/discovery/dns: add __meta_dns_srv_record_target and __meta_dns_srv_record_port labels
This syncs dns service discovery with Prometheus 2.21 - see https://github.com/prometheus/prometheus/releases
and https://github.com/prometheus/prometheus/pull/7678 .
2020-09-11 21:35:39 +03:00
Aliaksandr Valialkin
6382e8081a app/vmagent: allow setting multiple identical -remoteWrite.url values
This may be useful when each url is authenticated via different `-remoteWrite.basicAuth.username`.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/755
2020-09-11 15:17:29 +03:00
Aliaksandr Valialkin
5cf5a0e8c4 lib/protoparser/common: do not read request body when parsing timestamp query arg
This was preventing from reading data via /api/v1/prometheus/import .

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/750
2020-09-11 14:45:21 +03:00
Aliaksandr Valialkin
81c05f669b lib/storage: do not store inf values, since they may lead to significant precision loss for previously stored values
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/752
2020-09-11 14:45:20 +03:00
Aliaksandr Valialkin
8bd5aa3516 lib/protoparser: accept timestamp in milliseconds instead of seconds at /api/v1/import/prometheus
This improves consistency with timestamps in Prometheus text exposition format
2020-09-11 14:05:24 +03:00
Aliaksandr Valialkin
d3ad0d365e app/vmselect: move Deadline from netstorage to searchutils
This removes dependency on netstorage from searchutils.
2020-09-11 13:39:13 +03:00
Aliaksandr Valialkin
58d3b82ae5 app/{vminsert,vmagent}: allow passing timestamp via timestamp query arg when ingesting data to /api/v1/import/prometheus
See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/750
2020-09-11 13:28:31 +03:00
Nikolay Khramchikhin
af994562c8 Added endpointslices discovery to k8s api (#760)
This is similar to https://github.com/prometheus/prometheus/pull/6838 , which will be added in Prometheus v2.21.
See https://github.com/prometheus/prometheus/releases/tag/v2.21.0-rc.1

* Added endpointslices discovery to k8s api

Started from 1.17 k8s version endpointslices is beta,
it allows to query k8s api for endpoints more efficient.
It presents at scrape_config.yaml as separate role for kubernetes_sd_config.
kubernetes_sd_config:
- role: endpointslices

* fixed typos, changed EndpointConditions signature - with values instead of pointers
2020-09-11 12:24:50 +03:00
Aliaksandr Valialkin
579c20756a app/vmselect: substitute inf values at smooth_exponential with the previous values
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/757
2020-09-11 12:23:56 +03:00
Aliaksandr Valialkin
d67e6d3d2e app/vmselect: skip infinite values when calculating smooth_exponential
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/757
2020-09-11 11:57:53 +03:00
John Belmonte
a72a2566d7 fix typo on outliersk() doc (#758) 2020-09-11 01:59:55 +03:00
Aliaksandr Valialkin
06427a184f app/vmselect/graphite: typo fix in label name for vm_request_duration_seconds metric 2020-09-11 01:59:52 +03:00
Aliaksandr Valialkin
f307e6f432 app/vmselect: initial implementation of Graphite Metrics API
See https://graphite-api.readthedocs.io/en/latest/api.html#the-metrics-api
2020-09-11 00:30:20 +03:00
Aliaksandr Valialkin
87f916a2fb vendor: make vendor-update 2020-09-10 09:50:00 +03:00
Aliaksandr Valialkin
5afe69407d deployment/docker: update Go builder from v1.15.1 to v1.15.2
This fixes the following issues in Go runtime - see https://github.com/golang/go/issues?q=milestone%3AGo1.15.2+label%3ACherryPickApproved
2020-09-10 09:38:10 +03:00
Aliaksandr Valialkin
f5cb213ef9 lib/storage: reuse timestamp blocks for adjancent metric blocks with identical timestamps
This should reduce disk space usage when scraping targets containing metrics with identical names
such as `node_cpu_seconds_total`, histograms, quantiles, etc.

Expose `vm_timestamps_blocks_merged_total` and `vm_timestamps_bytes_saved_total` metrics for monitoring
the effectiveness of timestamp blocks merging.
2020-09-09 23:59:21 +03:00
Aliaksandr Valialkin
475698d2ad docs: sync docs for vmalert, vmauth, vmbackup and vmrestore 2020-09-09 21:10:48 +03:00
Aliaksandr Valialkin
d72fd86488 docs/Articles.md: add links to recently published third-party articles and talks about VictoriaMetrics 2020-09-09 20:15:44 +03:00
Aliaksandr Valialkin
5a82f645e4 docs/Single-server-VictoriaMetrics.md: typo fix 2020-09-09 01:22:18 +03:00
Aliaksandr Valialkin
95ad657a1c docs/Single-server-VictoriaMetrics.md: typo fix 2020-09-09 01:00:05 +03:00
Aliaksandr Valialkin
5ab57f916b docs/vmagent.md: clarified the case when -remoteWrite.queues must be tuned
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/745
2020-09-08 20:15:49 +03:00
Aliaksandr Valialkin
e5c8377212 lib/httpserver: add a jitter to connection timeouts in order to protect from Thundering herd problem 2020-09-08 19:57:20 +03:00
Aliaksandr Valialkin
df6519c190 vendor: make vendor-update 2020-09-08 15:18:59 +03:00
Aliaksandr Valialkin
f3a79abfb4 app/vmselect/promql: go fmt 2020-09-08 15:18:57 +03:00
Aliaksandr Valialkin
4f06eed1c1 app/vmselect/promql: adjust integrate() calculations to be more similar to calculations from InfluxDB: attempt #2
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/701
2020-09-08 14:36:23 +03:00
Aliaksandr Valialkin
0d0b606455 app/vmselect/promql: adjust integrate() calculations to be more similar to calculations from InfluxDB
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/701
2020-09-08 14:24:02 +03:00
Aliaksandr Valialkin
db91045348 app/vmselect/promql: increase floating point calculations accuracy by dividing by 1e3 instead of multiplying by 1e-3 2020-09-08 14:01:02 +03:00
Aliaksandr Valialkin
13dd915302 docs/Single-server-VictoriaMetrics.md: make docs-sync 2020-09-07 21:58:48 +03:00
Nikolay Khramchikhin
fb356c434b Changed s3 configProfile flag default, (#749)
aws sdk has complicated logic for chosing profile name and we shouldn't set
it to `default` value. It leads to bugs and improper configuration.
Set it to empty value by default is safe. It will be automatically set to `default` by sdk.
2020-09-07 21:55:45 +03:00
Aliaksandr Valialkin
5cb8c82fe5 docs/Single-server-VictoriaMetrics.md: sync with README.md 2020-09-04 03:30:43 +03:00
Aliaksandr Valialkin
38462bd95e docs/Single-server-VictoriaMetrics.md: updates according to review comments at fe98ba5a60 2020-09-04 03:08:20 +03:00
Aliaksandr Valialkin
804304c365 app/vmselect: add missing deletion for temporary files on partial responses when -search.denyPartialResponse=true 2020-09-04 02:23:12 +03:00
Aliaksandr Valialkin
d2f6f96e4a lib/memory: fall back to reading hierarchical memory limit in cgroups when the default limit isn't set
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/699
2020-09-04 00:04:57 +03:00