Commit graph

430 commits

Author SHA1 Message Date
Aliaksandr Valialkin
718eca33ab
lib/storage: properly handle {__name__=~"prefix(suffix1|suffix2)",other_label="..."} queries
They were broken in the commit 00cbb099b6

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1644
2021-09-23 21:48:51 +03:00
Aliaksandr Valialkin
8f685d81c6 lib/storage: follow up after 00cbb099b6 2021-09-14 14:16:25 +03:00
faceair
00cbb099b6
lib/storage: optimize convert multiple values regexp filter to composite tag filter (#1610)
* lib/storage: optimize convert multiple values regexp filter to composite tag filter

* Apply suggestions from code review

Co-authored-by: Aliaksandr Valialkin <valyala@gmail.com>
2021-09-14 12:47:07 +03:00
Aliaksandr Valialkin
cebcb15ba4 lib/storage: verify that the tsidsFound contain the needed tsids in tests added at f4dead529f 2021-09-11 10:57:13 +03:00
Aliaksandr Valialkin
f4dead529f lib/storage: properly search series by multiple tag filters matching empty labels such as foo{bar=~"baz|",x=~"y|"}
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1601
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/395
2021-09-09 21:09:21 +03:00
Aliaksandr Valialkin
f77dde837a lib/promscrape: add the ability to limit the number of unique series per each scrape target
The number of series per target can be limited with the following options:

* Global limit with `-promscrape.maxSeriesPerTarget` command-line option.
* Per-target limit with `max_series: N` option in `scrape_config` section.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1561
2021-09-01 16:03:59 +03:00
Aliaksandr Valialkin
ffc0ab1774 lib/{mergeset,storage}: improve the detection of the needed free space for background merge
This should prevent from possible out of disk space crashes during big merges.

See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1560
2021-08-25 09:35:44 +03:00
Aliaksandr Valialkin
4401464c22 all: add support for Prometheus staleness markers
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1526
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/748
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1509
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1530
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/845
2021-08-13 12:10:17 +03:00
Aliaksandr Valialkin
d05cac6c98 li/storage: re-use the per-day inverted index search code for searching in global index
This allows removing a big pile of outdated code for global index search.

This may help https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1486
2021-07-30 10:31:37 +03:00
Aliaksandr Valialkin
8ee8660ac4 app/vmselect: follow-up for 626073bca8
* Rename -search.maxMetricsPointSearch to -search.maxSamplesPerQuery, so it is more consistent with the existing -search.maxSamplesPerSeries
* Move the -search.maxSamplesPerQuery from vmstorage to vmselect, so it could effectively limit the number of raw samples obtained from all the vmstorage nodes
* Document the -search.maxSamplesPerQuery in docs/CHANGELOG.md
2021-07-28 18:00:23 +03:00
Aliaksandr Valialkin
c2deee9911 lib/storage: yet another attempt to properly determine disk space shortage, which prevents from optimal merges
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1373
2021-07-27 12:04:50 +03:00
Aliaksandr Valialkin
d472b03e34 lib/storage: make sure the second call to DeduplicateSamples and deduplicateSamplesDuringMerge doesnt change samples 2021-07-15 12:17:45 +03:00
Aliaksandr Valialkin
682662b2ae lib/storage: remove cache directory if it contains reset_cache_on_startup file
See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1447
2021-07-13 17:58:51 +03:00
Aliaksandr Valialkin
4f80b2f230 lib/storage: properly limit the size of storage/date_metricID cache 2021-07-12 14:25:44 +03:00
Aliaksandr Valialkin
8ca2799478 lib/storage: properly determine when the deduplication is needed in needsDedup
Previously needsDedup() could return true if the de-duplication wasn't needed for the following case:

         d < interval
           /     \
   |        v | v        |
     interval   interval

Now it properly returns false for this case
2021-07-12 10:53:30 +03:00
Aliaksandr Valialkin
8b262d4ba7 lib/storage: periodically reset prefetchedMetricIDs cache in order to limit its size under high churn rate 2021-07-07 10:58:51 +03:00
Aliaksandr Valialkin
a7694092b8 Revert "lib/uint64set: allow reusing bucket16 structs inside uint64set.Set via uint64set.Release method"
This reverts commit 7c6d3981bf.

Reason for revert: high contention at bucket16Pool on systems with big number of CPU cores.
This slows down query processing significantly.
2021-07-06 18:21:35 +03:00
Aliaksandr Valialkin
8aa9bba9bd lib/{mergeset,storage}: switch from sync.Pool to chan-based pool for inmemoryPart objects
This should reduce memory usage on systems with big number of CPU cores,
since every inmemoryPart object occupies at least 64KB of memory and sync.Pool maintains
a separate pool inmemoryPart objects per each CPU core.

Though the new scheme for the pool worsens per-cpu cache locality, this should be amortized
by big sizes of inmemoryPart objects.
2021-07-06 16:28:41 +03:00
Aliaksandr Valialkin
7c6d3981bf lib/uint64set: allow reusing bucket16 structs inside uint64set.Set via uint64set.Release method
This reduces the load on memory allocator in Go runtime in production workload.
2021-07-06 15:35:03 +03:00
Aliaksandr Valialkin
f3acf065c9 lib/storage: consistency renaming: tagCache -> tagFiltersCache
This improves code readability
2021-07-06 11:03:51 +03:00
Aliaksandr Valialkin
0020b9f904 lib/workingsetcache: properly update stats for requests and cache misses
Previously the stats for cache misses could be improperly counted, because it had inflated cache misses
if the entry was missing in the curr cache, but was existing in the prev cache.

The same applies to cache requests - they were inflated if the entry was missing in the curr cache.
2021-07-06 10:53:32 +03:00
Aliaksandr Valialkin
d0c830039d lib/storage: tune cache sizes according to production workload 2021-07-05 15:16:11 +03:00
Aliaksandr Valialkin
43103be011 lib/{storage,mergeset}: increase cache timeout for data and index blocks from a minute to two minutes
One minute cache timeout result in slower queries in some production workloads where the interval
between query execution is in the range 1 minute - 2 minutes.
2021-07-05 15:16:11 +03:00
Aliaksandr Valialkin
9a83e9018d lib/storage: properly detect free disk space shortage during data merge
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1373
2021-07-02 17:40:54 +03:00
Aliaksandr Valialkin
c93cee8de8 lib/{mergeset,storage}: reduce the maximum lifetime for cached indexdb and data blocks from 2 minutes to a minute
This should reduce memory usage on a system with high number of active time series and a high churn rate.
One minute is enough for caching the blocks needed for repeated queries (e.g. alerting rules, recording rules and dashboard refreshes).
2021-06-29 19:57:07 +03:00
Aliaksandr Valialkin
5506cff76e lib/storage: put indexDBName into the key for dateTagFilter cache and for uselessTagFilters cache
This should prevent from stats overwriting when the previous indexdb is queried.
2021-06-29 12:40:05 +03:00
Aliaksandr Valialkin
c22114c6f0 lib/storage: tune tag filters search logic
Tune the logic according to the logs provided at https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1338#issuecomment-864293624

The previous logic had a race when multiple concurrent queries execute the same tag filter without prior stats.
This could result in incorrectly stored stats for such tag filter, which then could result in non-optimal sorting of tag filters
for further queries.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1338
2021-06-23 13:29:39 +03:00
Aliaksandr Valialkin
dcbc22552f lib/storage: fix infinite loop introduced in aa9b56a046 2021-06-17 14:28:10 +03:00
Aliaksandr Valialkin
aa9b56a046 lib/{mergeset,storage}: reduce the number of fsync calls on data ingestion path on systems with many cpu cores
VictoriaMetrics maintains a buffer per CPU core for the ingested data. These buffers are flushed to disk every second.
These buffers are flushed to disk in parallel starting from the commit 56b6b893ce .
This resulted in increased write disk IO usage on systems with many cpu cores
as described at https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1338#issuecomment-863046999 .

This commit merges the per-CPU buffers into bigger in-memory buffers before flushing them to disk.
This should reduce the rate of fsync syscalls and, consequently, the write disk IO on systems with many CPU cores.

This should help https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1338
See also https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1244
2021-06-17 13:52:08 +03:00
Aliaksandr Valialkin
84fb59b0ba lib/storage: move deletedMetricIDs set from indexDB to Storage
This makes consitent the list of deleted metricIDs when it is used from both the current indexDB and the previous indexDB (aka extDB).
This should fix the issue, which could lead to storing new samples under deleted metricIDs after indexDB rotation.
See more details at https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1347#issuecomment-861232136 .

Thanks to @tangqipengleoo for the initial analysis and the pull request - https://github.com/VictoriaMetrics/VictoriaMetrics/pull/1383 .

This commit resolves the issue in more generic way compared to https://github.com/VictoriaMetrics/VictoriaMetrics/pull/1383 .

The downside of the commit is the deletedMetricIDs set isn't cleaned from the metricIDs outside the retention. It needs app restart.
This should be OK in most cases.
2021-06-15 15:04:30 +03:00
Aliaksandr Valialkin
c4f3fbfa5d lib/storage: reset cache on disk during series deletion and during indexdb rotation
This should prevent from inconsistent behavior (aka partially missing data for some time series) after unclean shutdown.

See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1347
2021-06-11 12:42:28 +03:00
Aliaksandr Valialkin
69b1482bdb lib/storage: consistency renaming: getMaxRawRowsPerPartition -> getMaxRawRowsPerShard 2021-06-11 10:57:23 +03:00
Aliaksandr Valialkin
044ab46824 lib/storage: reduce the amounts of memory which can be occupied by rawRow items during data ingestion on a system with many CPU cores 2021-06-11 10:57:23 +03:00
Aliaksandr Valialkin
96b691a0ab lib/storage: properly account the number of loops spent when matching for or suffixes
This may help https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1338
2021-06-08 13:06:12 +03:00
Aliaksandr Valialkin
2d8bd41f8a lib/storage: reduce memory allocations when syncing dateMetricIDCache 2021-06-03 16:20:42 +03:00
Aliaksandr Valialkin
39ef1e7a51 lib/storage: do not stop data ingestion on the first error in Storage.AddRows
Continue data ingestion for the rest of blocks.
2021-05-24 15:32:47 +03:00
Aliaksandr Valialkin
4b01c9fb2e lib/storage: limit the number of rows per each block in Storage.AddRows()
This should reduce memory usage when ingesting big blocks or rows.
2021-05-24 15:24:07 +03:00
Aliaksandr Valialkin
a4ff4b8e65 lib/storage: allow filling all the rows up to their capacity in rawRowsShard.addRows
This should reduce memory usage a bit on data ingestion path
2021-05-24 15:22:59 +03:00
Aliaksandr Valialkin
f54133b200 lib/storage: do not populate MetricID->MetricName cache during data ingestion
This cache isn't needed during data ingestion, so there is no need in spending RAM on it.

This reduces RAM usage on data ingestion path by 30%
2021-05-24 03:02:46 +03:00
Aliaksandr Valialkin
ec79abc382 lib/{mergeset,storage}: reduce the number of IFNO log messages like merged ... items across ... blocks in ... seconds
Log these messages if the merge takes more than 30 seconds instead of 10 seconds.
2021-05-23 14:03:21 +03:00
Aliaksandr Valialkin
ad73f226ff app/vmstorage: add ability to limit series cardinality via -storage.maxHourlySeries and -storage.maxDailySeries command-line flags 2021-05-20 14:15:19 +03:00
Aliaksandr Valialkin
009e136d88 lib/storage: remove possible data race when logging dropped labels 2021-05-20 02:47:22 +03:00
Aliaksandr Valialkin
d7be2753c0 lib/storage: substitute GetTSDBStatusForDate with GetTSDBStatusWithFiltersForDate with nil tfss 2021-05-13 09:02:33 +03:00
Aliaksandr Valialkin
a22a17dc66 lib/storage: merge getTSDBStatusForDate with getTSDBStatusWithFiltersForDate
These functions are non-trivial, while their code has minimal differences.
It is better from maintainability PoV to merge these functions into a single function.
2021-05-12 17:56:53 +03:00
Aliaksandr Valialkin
832651c6c2 app/vmselect: follow up after 8a0678678b
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1168
2021-05-12 17:18:30 +03:00
Nikolay
8a0678678b
Adds tsdb match filters (#1282)
* init work on filters

* init propose for status filters

* fixes tsdb status
adds test

* fix bug

* removes checks from test
2021-05-12 15:18:45 +03:00
Aliaksandr Valialkin
33f7bacb01 lib/storage: properly apply time range when matching an empty filter
It must match all the time series on the given time range.
Previously it was matched to all the time series without the restriction on the given time range.
2021-05-11 01:12:05 +03:00
Aliaksandr Valialkin
75c2c813fc lib/storage: remove dead code after the commit 3ccf7ea20c 2021-05-08 20:14:11 +03:00
Aliaksandr Valialkin
12d733dd5d app/vminsert: add support for data ingestion via other vminsert nodes 2021-05-08 19:52:57 +03:00
Aliaksandr Valialkin
76027093ca lib/storage: use WARNING instead of INFO level for logging dropped labels 2021-05-03 13:56:43 +03:00