Commit graph

5238 commits

Author SHA1 Message Date
Aliaksandr Valialkin
2db06f0ef8 lib/backup: prevent from restoring to directory which is in use by VictoriaMetrics during the restore 2019-11-19 18:36:23 +02:00
Aliaksandr Valialkin
0094bc4fc9 app/vmselect/prometheus: properly adjust too big time time on /api/v1/query
Too big `time` must be adjusted to `now()-queryOffset`.
2019-11-19 00:42:00 +02:00
Aliaksandr Valialkin
b6f22a62cb lib/storage: increase the number of created time series in BenchmarkHeadPostingForMatchers in order to be on par with Promethues
The previous commit was accidentally creating 10x smaller number of time series than Prometheus
and this led to invalid benchmark results.

The updated benchmark results:

benchmark                                                          old ns/op      new ns/op     delta
BenchmarkHeadPostingForMatchers/n="1"                              272756688      6194893       -97.73%
BenchmarkHeadPostingForMatchers/n="1",j="foo"                      138132923      10781372      -92.19%
BenchmarkHeadPostingForMatchers/j="foo",n="1"                      134723762      10632834      -92.11%
BenchmarkHeadPostingForMatchers/n="1",j!="foo"                     195823953      10679975      -94.55%
BenchmarkHeadPostingForMatchers/i=~".*"                            7962582919     100118510     -98.74%
BenchmarkHeadPostingForMatchers/i=~".+"                            7589543864     154955671     -97.96%
BenchmarkHeadPostingForMatchers/i=~""                              1142371741     258003769     -77.42%
BenchmarkHeadPostingForMatchers/i!=""                              9964150263     159783895     -98.40%
BenchmarkHeadPostingForMatchers/n="1",i=~".*",j="foo"              216995884      10937895      -94.96%
BenchmarkHeadPostingForMatchers/n="1",i=~".*",i!="2",j="foo"       202541348      10990027      -94.57%
BenchmarkHeadPostingForMatchers/n="1",i!=""                        486285711      87004349      -82.11%
BenchmarkHeadPostingForMatchers/n="1",i!="",j="foo"                350776931      53342793      -84.79%
BenchmarkHeadPostingForMatchers/n="1",i=~".+",j="foo"              380888565      54256156      -85.76%
BenchmarkHeadPostingForMatchers/n="1",i=~"1.+",j="foo"             89500296       21823279      -75.62%
BenchmarkHeadPostingForMatchers/n="1",i=~".+",i!="2",j="foo"       379529654      46671359      -87.70%
BenchmarkHeadPostingForMatchers/n="1",i=~".+",i!~"2.*",j="foo"     424563825      53915842      -87.30%

VictoriaMetrics uses 1GB of RAM during the benchmark (vs 3.5GB of RAM for Prometheus)
2019-11-18 19:50:58 +02:00
Aliaksandr Valialkin
8a0dfc6220 lib/storage: add BenchmarkHeadPostingForMatchers similar to the benchmark from Prometheus
See the corresponding benchmark in Prometheus - 23c0299d85/tsdb/head_bench_test.go (L52)

The benchmark allows performing apples-to-apples comparison of time series search
in Prometheus and VictoriaMetrics. The following article - https://www.robustperception.io/evaluating-performance-and-correctness -
contains incorrect numbers for VictoriaMetrics, since there wasn't this benchmark yet. Fix this.

Benchmarks can be repeated with the following commands from Prometheus and VictoriaMetrics source code roots:

- Prometheus: GOMAXPROCS=1 go test ./tsdb/ -run=111 -bench=BenchmarkHeadPostingForMatchers
- VictoriaMetrics: GOMAXPROCS=1 go test ./lib/storage/ -run=111 -bench=BenchmarkHeadPostingForMatchers

Benchmark results:
benchmark                                                          old ns/op      new ns/op     delta
BenchmarkHeadPostingForMatchers/n="1"                              272756688      364977        -99.87%
BenchmarkHeadPostingForMatchers/n="1",j="foo"                      138132923      1181636       -99.14%
BenchmarkHeadPostingForMatchers/j="foo",n="1"                      134723762      1141578       -99.15%
BenchmarkHeadPostingForMatchers/n="1",j!="foo"                     195823953      1148056       -99.41%
BenchmarkHeadPostingForMatchers/i=~".*"                            7962582919     8716755       -99.89%
BenchmarkHeadPostingForMatchers/i=~".+"                            7589543864     12096587      -99.84%
BenchmarkHeadPostingForMatchers/i=~""                              1142371741     16164560      -98.59%
BenchmarkHeadPostingForMatchers/i!=""                              9964150263     12230021      -99.88%
BenchmarkHeadPostingForMatchers/n="1",i=~".*",j="foo"              216995884      1173476       -99.46%
BenchmarkHeadPostingForMatchers/n="1",i=~".*",i!="2",j="foo"       202541348      1299743       -99.36%
BenchmarkHeadPostingForMatchers/n="1",i!=""                        486285711      11555193      -97.62%
BenchmarkHeadPostingForMatchers/n="1",i!="",j="foo"                350776931      5607506       -98.40%
BenchmarkHeadPostingForMatchers/n="1",i=~".+",j="foo"              380888565      6380335       -98.32%
BenchmarkHeadPostingForMatchers/n="1",i=~"1.+",j="foo"             89500296       2078970       -97.68%
BenchmarkHeadPostingForMatchers/n="1",i=~".+",i!="2",j="foo"       379529654      6561368       -98.27%
BenchmarkHeadPostingForMatchers/n="1",i=~".+",i!~"2.*",j="foo"     424563825      6757132       -98.41%

The first column (old) is for Prometheus, the second column (new) is for VictoriaMetrics.

As you can see, VictoriaMetrics outperforms Prometheus by more than 100x in almost all the test cases of this benchmark.

Prometheus was using 3.5GB of RAM during the benchmark, while VictoriaMetrics was using 400MB of RAM.
2019-11-18 18:45:06 +02:00
Aliaksandr Valialkin
2ab4cea5e5 lib/storage: always start using per-day inverted index on the next day after its creation
The current day could miss entries for already stopped time series before
enabling per-day index.

This fixes the issue when queries return empty results during the first hour after
upgrading to v1.29.*
2019-11-16 12:11:25 +02:00
Aliaksandr Valialkin
c050abbbad deployment/docker: update Prometheus version from v2.12.0 to v2.14.0 2019-11-16 00:13:15 +02:00
Aliaksandr Valialkin
3f1637fae8 app/vmselect/promql: properly calculate integrate(q[d]) 2019-11-13 21:10:41 +02:00
Aliaksandr Valialkin
c56b9ed03b app/victoria-metrics: add build rules for GOARCH=ppc64le
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/235
2019-11-13 20:24:33 +02:00
Aliaksandr Valialkin
3fd32e331a app/vmselect/promql: use universal approach for determining maxByteSliceLen on 32-bit and 64-bit archs
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/235
2019-11-13 20:24:26 +02:00
Aliaksandr Valialkin
119dfd01bb lib/storage: add vm_cache_size_bytes{type="storage/hour_metric_ids"} metric 2019-11-13 20:24:21 +02:00
Aliaksandr Valialkin
86a1cd700b lib/storage: remove inmemory index for recent hour, since it uses too much memory
Production workload shows that the index requires ~4Kb of RAM per active time series.
This is too much for high number of active time series, so let's delete this index.

Now the queries should fall back to the index for the current day instead of the index
for the recent hour. The query performance for the current day index should be good enough
given the 100M rows/sec scan speed per CPU core.
2019-11-13 17:58:07 +02:00
Aliaksandr Valialkin
33895d4a0f lib/storage: add missing increment for recentHourInvertedIndexSearchCalls 2019-11-13 15:13:51 +02:00
Aliaksandr Valialkin
c57eb0ff83 lib/storage: add -disableRecentHourIndex flag for disabling inmemory index for recent hour
This may be useful for saving RAM on high number of time series aka high cardinality
2019-11-13 15:02:51 +02:00
Aliaksandr Valialkin
e14ab14e54 lib/storage: verify marshaling for iidx.pendingMetricIDs in TestInmemoryInvertedIndexMarshalUnmarshal 2019-11-13 13:35:30 +02:00
Aliaksandr Valialkin
ca259864e2 lib/storage: return back inmemory inverted index for recent hour
Issues fixed:
- Slow startup times. Now the index is loaded from cache during start.
- High memory usage related to superflouos index copies every 10 seconds.
2019-11-13 13:11:04 +02:00
Aliaksandr Valialkin
01bb3c06c7 lib/storage: remove inmemory inverted index for recent hours
Production load with >10M active time series showed it could
slow down VictoriaMetrics startup times and could eat
all the memory leading to OOM.

Remove inmemory inverted index for recent hours until thorough
testing on production data shows it works OK.
2019-11-13 10:45:53 +02:00
Aliaksandr Valialkin
66c4961ff8 README.md: mention that VictoriaMetrics executable is small 2019-11-12 16:58:15 +02:00
Aliaksandr Valialkin
3e16248ed6 README.md: small updates 2019-11-12 16:54:18 +02:00
Aliaksandr Valialkin
5e6c1cd986 README.md: typo fix 2019-11-12 16:48:40 +02:00
Aliaksandr Valialkin
6c2303764e Revert "lib/fs: do not postpone directory removal on NFS error"
This reverts commit 4c02e496f7.

Reason for revert: the commit breaks on NFS - see https://github.com/VictoriaMetrics/VictoriaMetrics/issues/234
2019-11-12 16:18:09 +02:00
Mike Poindexter
f3ad330635 Add test for invalid caching of tsids (#232)
* Add test for invalid caching of tsids

* Clean up error handling
2019-11-12 15:09:33 +02:00
Aliaksandr Valialkin
6c362d82cb README.md: mention that backups are made to S3 or GCS 2019-11-12 14:32:37 +02:00
Aliaksandr Valialkin
661dd190bb Refer to https://medium.com/@valyala/speeding-up-backups-for-big-time-series-databases-533c1a927883 from multiple places in README.md 2019-11-12 13:02:39 +02:00
Aliaksandr Valialkin
630ba810f1 deployment/docker: upgrade Go from v1.13.4 to v1.13.4 2019-11-12 03:49:19 +02:00
Oleg Kovalov
b4f44befa3 fix misspelled words (#229) 2019-11-12 00:16:42 +02:00
Roman Khavronenko
5fc8fb1323 add churn rate panel (#230) 2019-11-12 00:14:53 +02:00
Aliaksandr Valialkin
8e8f98f712 lib/storage: add tests for dateMetricIDCache 2019-11-11 13:21:57 +02:00
Aliaksandr Valialkin
c342f5e37e lib/storage: eliminate data race when updating lastSyncTime in dateMetricIDCache.Has 2019-11-10 22:04:01 +02:00
Aliaksandr Valialkin
56d7cc8a0d app/victoria-metrics: remove deprecated fs.MustStopDirRemover from main_test.go 2019-11-10 13:37:13 +02:00
Aliaksandr Valialkin
4c02e496f7 lib/fs: do not postpone directory removal on NFS error
Continue trying to remove NFS directory on temporary errors for up to a minute.

The previous async removal process breaks in the following case during VictoriaMetrics start

- VictoriaMetrics opens index, finds incomplete merge transactions and starts replaying them.
- The transaction instructs removing old directories for parts, which were already merged into bigger part.
- VictoriaMetrics removes these directories, but their removal is delayed due to NFS errors.
- VictoriaMetrics scans partition directory after all the incomplete merge transactions are finished
  and finds directories, which should be removed, but weren't still removed due to NFS errors.
- VictoriaMetrics panics when it finds unexpected empty directory.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/162
2019-11-10 13:24:51 +02:00
Aliaksandr Valialkin
3956003dd0 lib/storage: reorganize the code in getStartDateForPerDayInvertedIndex according to golangci-lint 2019-11-10 00:38:59 +02:00
Aliaksandr Valialkin
5c3fa59181 app/vmrestore: the upcoming release would be 1.29.0 2019-11-10 00:20:41 +02:00
Aliaksandr Valialkin
ee7765b10d lib/storage: implement per-day inverted index 2019-11-10 00:02:46 +02:00
Aliaksandr Valialkin
5810ba57c2 lib/storage: use specialized cache for (date, metricID) entries
This improves ingestion performance.
2019-11-09 23:06:11 +02:00
Aliaksandr Valialkin
e573ef2126 lib/storage: remove unused code from getMetricIDsForTimeRange: it is expected that time range is always non-zero 2019-11-09 19:03:34 +02:00
Aliaksandr Valialkin
823fa085ef lib/storage: properly set time range when deleting time series 2019-11-09 18:49:49 +02:00
Aliaksandr Valialkin
695c1dc5eb lib/storage: obtain all the time series ids from (tag->metricIDs) rows instead of (metricID->TSID) rows, since this much faster 2019-11-09 18:04:33 +02:00
Aliaksandr Valialkin
cdbe848102 lib/storage: small code prettifying 2019-11-09 14:19:52 +02:00
Aliaksandr Valialkin
5c25070556 lib/uint64set: remove superflouos check for item existence before deleting it in Set.Subtract 2019-11-09 14:19:47 +02:00
Aliaksandr Valialkin
bb08bab263 lib/storage: inmemoryInvertedIndex prettifying 2019-11-09 14:19:41 +02:00
Aliaksandr Valialkin
6ad7fe8eeb lib/storage: export vm_new_timeseries_created_total metric for determining time series churn rate 2019-11-08 21:21:07 +02:00
Aliaksandr Valialkin
9ea549ed24 lib/storage: sync with cluster changes 2019-11-08 21:21:07 +02:00
Aliaksandr Valialkin
63b05c0b9f app/vmselect/promql: adjust memory limits calculations for incremental aggregate functions
Incremental aggregate functions don't keep all the selected time series in memory -
they keep only up to GOMAXPROCS time series for incremental aggregations.

Take into account that the number of time series in RAM can be higher if they are split
into many groups with `by (...)` or `without (...)` modifiers.

This should reduce the number of `not enough memory for processing ... data points` false
positive errors.
2019-11-08 21:21:07 +02:00
Aliaksandr Valialkin
d888b21657 lib/storage: add inmemory inverted index for the last hour
It should improve performance for `last N hours` dashboards with update intervals smaller than 1 hour.
2019-11-08 21:21:07 +02:00
Aliaksandr Valialkin
1e46961d68 app/{vmbackup,vmrestore}: add vmbackup and vmrestore tools for creating backups on s3 or gcs from instant snapshots
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/203
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/38
2019-11-08 21:21:07 +02:00
Roman Khavronenko
72756ab8c7 #224: add slow_queries, on-going merges and merge speed panels to dashboard (#226) 2019-11-08 21:20:38 +02:00
Aliaksandr Valialkin
543dc8d337 lib/storage: populate partition names from both small and big directories
Certain partition directories may be missing after restoring from backups
if they had no data. Re-create such directories on start.
2019-11-06 19:49:34 +02:00
Aliaksandr Valialkin
e472f0b23b lib/storage: substitute error message about unsorted items in the index block after metricIDs merge with counter
The origin of the error has been detected and documented in the code,
so it is enough to export a counter for such errors at `vm_index_blocks_with_metric_ids_incorrect_order_total`,
so it could be monitored and alerted on high error rates.

Export also the counter for processed index blocks with metricIDs - `vm_index_blocks_with_metric_ids_processed_total`,
so its' rate could be compared to `rate(vm_index_blocks_with_metric_ids_incorrect_order_total)`.
2019-11-06 14:28:11 +02:00
Aliaksandr Valialkin
c51ca04a43 lib/storage: take into account the requested time range when caching TSIDs for the given tag filters 2019-11-06 14:28:11 +02:00
Aliaksandr Valialkin
e37f06dc52 lib/storage: dump incorrectly sorted items on a single line; this should simplify error reporting 2019-11-05 18:44:22 +02:00