prometheus

Commit Graph

Author	SHA1	Message	Date
beorn7	c5fa0b90c3	Fix the case where a series in memory has 0 chunks, but chunks on disk. This is actually completely normal for a freshly unarchived series. Test added to expose.	10 years ago
beorn7	3035b8bfdd	Adaptively reduce the wait time for memory series maintenance. This will make in-memory series maintenance the faster the more chunks are waiting for persistence.	10 years ago
beorn7	6a21f73898	Fixes after review.	10 years ago
beorn7	51d35f4481	Instrument series maintenance durations.	10 years ago
beorn7	12ae6e9203	Increase resilience of the storage against data corruption - step 4. Step 4: Add a configurable sync'ing of series files after modification.	10 years ago
beorn7	11bd9ce1bd	Increase resilience of the storage against data corruption - step 3. Step 3: Remember the mtime of series files and make use of it to detect series files that are not the one the checkpoint thinks they are.	10 years ago
beorn7	e25cca823c	Increase resilience of the storage against data corruption - step 2. Step 2: Add a flag -storage.local.pedantic-checks to check every series file. Also, remove countPersistedHeadChunks channel, which is unused.	10 years ago
beorn7	3d8d8928be	Increase resilience of the storage against data corruption - step 1. Step 1: Admit the problem by turning the various "panic"s into logged errors, followed by marking the persistence as dirty.	10 years ago
beorn7	da7c0461c6	Rename persist queue len/cap to num/max chunks to persist. Remove deprecated flag storage.incoming-samples-queue-capacity.	10 years ago
beorn7	a075900f9a	Merge branch 'beorn7/persistence' into beorn7/ingestion-tweaks	10 years ago
beorn7	1d8fc7d56f	Change minor things after code review.	10 years ago
beorn7	be11cb2b07	Remove the sample ingestion channel. The one central sample ingestion channel has caused a variety of trouble. This commit removes it. Targets and rule evaluation call an Append method directly now. To incorporate multiple storage backends (like OpenTSDB), storage.Tee forks the Append into two different appenders. Note that the tsdb queue manager had its own queue anyway. It was a queue after a queue... Much queue, so overhead... Targets have their own little buffer (implemented as a channel) to avoid stalling during an http scrape. But a new scrape will only be started once the old one is fully ingested. The contraption of three pipelined ingesters was removed. A Target is an ingester itself now. Despite more logic in Target, things should be less confusing now. Also, remove lint and vet warnings in ast.go.	10 years ago
beorn7	0056eaeb4f	Redesign series maintenance and chunk persistence.	10 years ago
beorn7	5bea942d8e	Improve various things around chunk encoding. A number of mostly minor things: - Rename chunk type -> chunk encoding. - After all, do not carry around the chunk encoding to all parts of the system, but just have one place where the encoding for new chunks is set based on the flag. The new approach has caveats as well, but the polution of so many method signatures is worse. - Use the default chunk encoding for new chunks of existing series. (Previously, only new _series_ would get chunks with the default encoding.) - Use an enum for chunk encoding. (But keep the version number for the flag, for reasons discussed previously.) - Add encoding() to the chunk interface (so that a chunk knows its own encoding - no need to have that in a different top-level function). - Got rid of newFollowUpChunk (which would keep the existing encoding for all chunks of a time series). Now only use newChunk(), which will create a chunk encoding according to the flag. - Simplified transcodeAndAdd. - Reordered methods of deltaEncodedChunk and doubleDeltaEncoded chunk to match the order in the chunk interface. - Only transcode if the chunk is not yet half full. If more than half full, add a new chunk instead.	10 years ago
beorn7	13fcf1ddbc	Implement double-delta encoded chunks.	10 years ago
beorn7	5ed8f6c205	Update persistQueueLength after chunks were persisted.	10 years ago
beorn7	1db7589081	Reduce the capacity of countPersistedHeadChunks. The capacity is basically how many persisted head chunks we will count at most while doing other things, in particular checkpointing. To limit the amount of already counted head chunks, keep this number low, otherwise we will easily checkpoint too often if checkpoints take long anyway.	10 years ago
beorn7	9406afad72	Do not double-count non-persisted head chunks on loading.	10 years ago
beorn7	dbc22b972c	Check last time in head chunk for head chunk timeout, not first.	10 years ago
beorn7	edd716e63c	Fix the embarrassing bug introduced in commit `0851945`. In that commit, the 'maintainSeries' call was accidentally removed. This commit refactors things a bit so that there is now a clean 'maintainMemorySeries' and a 'maintainArchivedSeries' call. Straighten the nomenclature a bit (consistently use 'drop' for chunks and 'purge' for series/metrics). Remove the annoying 'Completed maintenance sweep through archived fingerprints' message if there were no archived fingerprints to do maintenance on.	10 years ago
beorn7	af91fb8e31	Improve persisting chunks to disk. This is done by bucketing chunks by fingerprint. If the persisting to disk falls behind, more and more chunks are in the queue. As soon as there are "double hits", we will now persist both chunks in one go, doubling the disk throughput (assuming it is limited by disk seeks). Should even more pile up so that we end wit "triple hits", we will persist those first, and so on. Even if we have millions of time series, this will still help, assuming not all of them are growing with the same speed. Series that get many samples and/or are not very compressable will accumulate chunks faster, and they will soon get double- or triple-writes. To improve the chance of double writes, -storage.local.persistence-queue-capacity could be set to a higher value. However, that will slow down shutdown a lot (as the queue has to be worked through). So we leave it to the user to set it to a really high value. A more fundamental solution would be to checkpoint not only head chunks, but also chunks still in the persist queue. That would be quite complicated for a rather limited use-case (running many time series with high ingestion rate on slow spinning disks).	10 years ago
beorn7	e22f26bc58	Move to a queue model for appending samples after all. Starting a goroutine takes 1-2µs on my laptop. From the "numbers every Go programmer should know", I had 300ns for a channel send in my mind. Turns out, on my laptop, it takes only 60ns. That's fast enough to warrant the machinery of yet another channel with a fixed set of worker goroutines feeding from it. The number chosen (8 for now) is low enough to not really afflict a measurable overhead (a big Prometheus server has >1000 goroutines running), but high enough to not make sample ingestion a bottleneck.	10 years ago
beorn7	fe518fdb28	Simplify AppendSamples by allowing it to be goroutine-unsafe.	10 years ago
beorn7	5d3cd65a5d	Improve performance of ingestion. - Parallelize AppendSamples as much as possible without breaking the contract about temporal order. - Allocate more fingerprint locker slots. - Do not run early checkpoints if we are behind on chunk persistence. - Increase fpMinWaitDuration to give the disk more time for more important things. Also, switch math.MaxInt64 and math.MinInt64 to the new constants.	10 years ago
beorn7	d2ab49c396	Make the persist queue length configurable. Also, set a much higher default value. Chunk persist requests can be quite spiky. If you collect a large number of time series that are very similar, they will tend to finish up a chunk at about the same time. There is no reason we need to back up scraping just because of that. The rationale of the new default value is "1/8 of the chunks in memory".	10 years ago
Julius Volz	9412b296d5	Remove labels on persist error counter. This fixes https://github.com/prometheus/prometheus/issues/496	10 years ago
Bjoern Rabenstein	2c8fdcbc23	Remove a deadlock during shutdown. If queries are still running when the shutdown is initiated, they will finish _during_ the shutdown. In that case, they might request chunk eviction upon unpinning their pinned chunks. That might completely fill the evict request queue _after_ draining it during storage shutdown. If that ever happens (which is the case if there are _many_ queries still running during shutdown), the affected queries will be stuck while keeping a fingerprint locked. The checkpointing can then not process that fingerprint (or one that shares the same lock). And then we are deadlocked.	10 years ago
Bjoern Rabenstein	5859b74f1b	Clean up license issues. - Move CONTRIBUTORS.md to the more common AUTHORS. - Added the required NOTICE file. - Changed "Prometheus Team" to "The Prometheus Authors". - Reverted the erroneous changes to the Apache License.	10 years ago
Julius Volz	a6bc42bc61	Minor formatting/spelling fixups.	10 years ago
Bjoern Rabenstein	0851945054	Add a heuristics to checkpoint early if there are many "dirty" series..	10 years ago
Bjoern Rabenstein	622e8350cd	Fix a bug handling freshly unarchived series. Usually, if you unarchive a series, it is to add something to it, which will create a new head chunk. However, if a series in unarchived, and before anything is added to it, it is handled by the maintenance loop, it will be archived again. In that case, we have to load the chunkDescs to know the lastTime of the series to be archived. Usually, this case will happen only rarely (as a race, has never happened so far, possibly because the locking around unarchiving and the subsequent sample append is smart enough). However, during crash recovery, we sometimes treat series as "freshly unarchived" without directly appending a sample. We might add more cases of that type later, so better deal with archiving properly and load chunkDescs if required.	10 years ago
Bjoern Rabenstein	eb932d1524	Remove a deadlock during shutdown.	10 years ago
Brian Brazil	e56786b221	Have scrape time as a pseudovariable, not a prometheus variable. This ensures it has the right timestamp, and is easier to work with. Switch sd variable away from 'outcome', using total/failed instead.	10 years ago
Julius Volz	c9618d11e8	Introduce copy-on-write for metrics in AST. This depends on changes in: https://github.com/prometheus/client_golang/tree/cow-metrics. Change-Id: I80b94833a60ddf954c7cd92fd2cfbebd8dd46142	10 years ago
Bjoern Rabenstein	674624f1c8	Completed more TODOs. - Documented checkpoint file format. - High-level description of series sanitation. - Replace fp.LoadFromString panic with an error. (Change in client_golang already submitted.) - Introduced checks for series file size where appropriate. - Removed two Law of Demeter violations. Change-Id: I555d97a2c8f4769820c2fc8bf5d6f4e160222abc	10 years ago
Bjoern Rabenstein	7d11019aa2	Squash a few trivial TODOs. - Delete unneeded file view_adapter.go. - Assessed that we still need the fingerprints in nodes (to create iterators). - Turned numMemChunkDescs into a metric. Change-Id: I29be963c795a075ec00c095f76bf26405535609d	10 years ago
Bjoern Rabenstein	14bda4180c	Changes after pair code review. Change-Id: Ib72d40f8e9027818cfbbd32a7a7201eebda07455	10 years ago
Bjoern Rabenstein	9ea808cd8b	Remove debug log line. Change-Id: Icdd2351b89f2d37ac2b615f9cf872e054c694ad1	10 years ago
Bjoern Rabenstein	bb42cc2e2d	Evict based on memory pressure. Evict recently used chunks last. Change-Id: Ie6168f0cdb3917bdc63b6fe15585dd70c1e42afe	10 years ago
Bjoern Rabenstein	d73e851b14	Tweak timing in the maintenance loop. Change-Id: I9801c4f9a22c3b3dc1ce1af81fdd9e992a4f4dd7	10 years ago
Bjoern Rabenstein	2672aa8ece	Instrument series maintenance. Change-Id: Ie4269d07ad4d23d44230c95a523088b472718e54	10 years ago
Bjoern Rabenstein	74c143c4c9	Improve scraper shutdown time. - Stop target pools in parallel. - Stop individual scrapers in goroutines, too. - Timing tweaks. Change-Id: I9dff1ee18616694f14b04408eaf1625d0f989696	10 years ago
Bjoern Rabenstein	3f61d304ce	Reorganize maintenance loop. Change-Id: Iac10f988ba3e93ffb188f49c30f92e0b6adce5a3	10 years ago
Bjoern Rabenstein	c087ee35f7	Remove archiveMtx. Change-Id: Ie8019f860bbda68621f74380c90a4e57930d3d7a	10 years ago
Bjoern Rabenstein	7af42eda65	Optimize purging. Now only purge if there is something to purge. Also, set savedFirstTime and archived time range appropriately. (Which is needed for the optimization.) Change-Id: Idcd33319a84def3ce0318d886f10c6800369e7f9	10 years ago
Bjoern Rabenstein	904acd43da	Add crash recovery. Fix the behavior if preload for non-existent series is requested. Instead of returning an error (which triggers a panic further up), simply count those incidents. They can happen regularly, we just want to know if they happen too frequently because that would mean the indexing is behind or broken. Change-Id: I4b2d1b93c4146eeea897d188063cb9574a270f8b	10 years ago
Bjoern Rabenstein	4efc60174b	Tweak and verify a few parameters. Remove TODOs accordingly. Change-Id: Ic062e13b6ae89a9135d3f14011114fe1cca1cef8	10 years ago
Bjoern Rabenstein	5f8e9617ef	Add more tests. Add an end-to-end fuzz and race test. Fix a race exposed by the above. Change-Id: Ifaa39a90cefbde8d4c29bda197cc92592ded21bb	10 years ago
Bjoern Rabenstein	d215e013b7	Fix the weird chunkDesc shuffling bug. The root cause was that after chunkDesc eviction, the offset between memory representation of chunk layout (via chunkDescs in memory) was shiftet against chunks as layed out on disk. Keeping the offset up to date is by no means trivial, so this commit is pretty involved. Also, found a race that for some reason didn't bite us so far: Persisting chunks was completel unlocked, so if chunks were purged on disk at the same time, disaster would strike. However, locking the persisting of chunk revealed interesting dead locks. Basically, never queue under the fp lock. Change-Id: I1ea9e4e71024cabbc1f9601b28e74db0c5c55db8	10 years ago
Bjoern Rabenstein	f1de5b0c4e	Run checkpointing of in-memory metrics and head chunks periodically. Checkpointing interval is now a command line flag. Along the way, several things were refactored. - Restructure the way the storage is started and stopped.. - Number of series in checkpoint is now a uint64, not a varint. (Breaks old checkpoints, needs wipe!) - More consistent naming and order of methods. Change-Id: I883d9170c9a608ee716bb0ab3d0ded8ca03760d9	10 years ago

1 2

83 Commits (f81cbb00c623a14ae750e1ee0751f0ba2cf2e4e6)