118 Commits (5579efbd5bb3a478ae1d849ab3ae6024b4ab43eb)

Author	SHA1	Message	Date
Fabian Reinartz	30efe4a58c	Support writing to multiple head blocks ... This is an initial (and hacky) first pass on allowing appending to multiple blocks simultaniously to avoid dropping samples right after cutting a new head block. It's also required for cases like the PGW, where a scrape may contain varying timestamps.	8 years ago
Fabian Reinartz	c20cc44b06	Add docs, write sequence number to meta.json	8 years ago
Fabian Reinartz	472c618c39	Drop out-of-bound samples	8 years ago
Fabian Reinartz	d4779b374c	Properly track and write meta file	8 years ago
Fabian Reinartz	9ddbd64d00	Move stats into meta.json file, cleanup, docs	8 years ago
Fabian Reinartz	5ceca3c810	Write to WAL before appending to memory storage	8 years ago
Fabian Reinartz	343dd9d94c	Fix wrong byte size in WAL base ref	8 years ago
Fabian Reinartz	5fb01d41aa	Use new Prometheus text format parser	8 years ago
Fabian Reinartz	dd0b69fe1b	Export ErrNotFound	8 years ago
Fabian Reinartz	1c80c33e72	Fix bug of unsorted postings lists being created ... The former approach created unordered postings list by either map iteration of new series being unsorted (fixable) or concurrent writers creating new series interleaved. We switch back to generating ephemeral references for a single batch. Newly created series have to be re-set upon the next insert.	8 years ago
Fabian Reinartz	c7f5590a71	Ensure order of postings when adding new series	8 years ago
Fabian Reinartz	d970f0256a	Add Rollback() and docs to Appender interface	8 years ago
Fabian Reinartz	22db9c3413	Remove old appendBatch methods	8 years ago
Fabian Reinartz	fde69dab49	Use buffer pool for head appenders	8 years ago
Fabian Reinartz	a317f252b9	Expose series references to clients ... This exposes a reference number of a series represented by a label set to clients. Subsequent samples can be directly added via the reference rather than repeatedly passing in the full labels. This drasitcally speeds up the append process. The appender chain uses different sections of the reference number for assignment to child appenders and invalidating reference numbers as necessary. Clients can either pass out reference numbers themselves or have their own optimized lookup, i.e. by directly associating unparsed metric descriptors strings with reference numbers.	8 years ago
Fabian Reinartz	5e028710d5	Add fast past to validation after lock switch	8 years ago
Fabian Reinartz	1b39887baa	Revalidate series existance after lock switch	8 years ago
Fabian Reinartz	ca5791efbc	Simplify creation of new series	8 years ago
Fabian Reinartz	0ca755b4ae	Replace single head chunk per series with memSeries ... This adds a memory series holding several chunk to replace the single head chunk per series so far. This is necessary for uniform maximum chunk sizes in cases where some series have higher frequency samples than others.	8 years ago
Fabian Reinartz	c32a94d409	Unexport HeadBlock, export Block interface	8 years ago
Fabian Reinartz	d86e8a63c7	Report correct number of appended samples	8 years ago
Fabian Reinartz	29883a18fc	Add own Appender() method for DB	8 years ago
Fabian Reinartz	4c4e0c614e	Simplify position mapper updating	8 years ago
Fabian Reinartz	0dffd52238	Use page writer in compaction	8 years ago
Fabian Reinartz	89d8467f5c	Add missing lock	8 years ago
Fabian Reinartz	8c31c6e934	Make concurrent head chunk reads safe, fix misc races ... This adds a 4 sample buffer to every head chunk. The XOR compression scheme may edit bytes in place. The minimum size of a sample is 2 bits. So keeping the last 4 samples in an in-memory buffer makes it safe to query the preceeding ones while samples are added	8 years ago
Fabian Reinartz	1943f8d1bb	Fix head block stats races	8 years ago
Fabian Reinartz	6aa922c5a6	Fix races	8 years ago
Fabian Reinartz	300f4e2abf	Use separate lock for series creation ... This uses the head block's own lock to only lock if new series were encountered. In the general append case we just need to hold a	8 years ago
Fabian Reinartz	63e12807da	Don't update head postings mapper on every append	8 years ago
Fabian Reinartz	71efd2e08d	Periodically fsync WAL, make head cut async	8 years ago
Fabian Reinartz	96c2bd249f	Handle compaction trigger and reinitializing in DB	8 years ago
Fabian Reinartz	937cdb579c	Switch to sequential block names ... This changes block directory names from the int64 timestamp to sequential numbering.	8 years ago
Fabian Reinartz	9790aa98ac	Add postings wrapper that emits head postings in label set order ... This adds a position mapper that takes series from a head block in the order they were appended and creates a mapping representing them in order of their label sets. Write-repair of the postings list would cause very expensive writing. Hence, we keep them as they are and only apply the postition mapping at the very end, after a postings list has been sufficienctly reduced through intersections etc.	8 years ago
Fabian Reinartz	5aa7f7cce8	Compact head block into persisted block	8 years ago
Fabian Reinartz	3f72d5d027	Fix last timestamp initialization ... This initializes the chunkDesc's last timestamp to the minimum value so initial samples with a timestamp of 0 (e.g. in tests) are not accidentally dropped.	8 years ago
Fabian Reinartz	ac49f8c15e	Consolidate persistence and compaction	8 years ago
Fabian Reinartz	e7f04d14d5	Lock mmapped files	8 years ago
Fabian Reinartz	91b65b55e7	Run persistence in separate goroutine	8 years ago
Fabian Reinartz	a648ef5252	Convert persister into function	8 years ago
Fabian Reinartz	dbd2b21d2e	Make persistence atomic	8 years ago
Fabian Reinartz	beb842a856	Change block interface ... This changes the block interface to directly expose index and series readers rather than the direct querier	8 years ago
Fabian Reinartz	5d7ec06e04	Use ChunkMeta in SeriesWriter	8 years ago
Fabian Reinartz	c00d17e691	Modify IndexReader API to accomodate compaction ... This changes the IndexReader API to expose plain labels and chunk meta information instead of a Series interface. Dropping of irrelevant chunks is moved into the querier. A LabelIndices method is added to query for existing label value indices.	8 years ago
Fabian Reinartz	1e1a37b15b	Remove double-reference in chunk hashmap	8 years ago
Fabian Reinartz	675f0886f0	Append to chunks cannot error	8 years ago
Fabian Reinartz	7280533c42	Add basic shard metrics	8 years ago
Fabian Reinartz	201d7687b6	Fix uniqueness of new series	8 years ago
Fabian Reinartz	787199a88e	Fix erroneous value assignments	8 years ago
Fabian Reinartz	8aba95048a	Rename to OpenHeadBlock	8 years ago
Fabian Reinartz	1dde3b6d31	Add WAL decoder+loading and benchmarks	8 years ago
Fabian Reinartz	0b8c77361e	Add initial WAL writing	8 years ago
Fabian Reinartz	2a825f6c28	Consolidate mem index into HeadBlock	8 years ago
Fabian Reinartz	ede733ab6c	Extract labels package	8 years ago
Fabian Reinartz	1b23d62e3f	Properly close files before reopening	8 years ago
Fabian Reinartz	282d9ae6e2	Implement label value queries in all layers.	8 years ago
Fabian Reinartz	bad93d8d57	Extract head serialization into Head method	8 years ago
Fabian Reinartz	b08f82fa4e	Pre-select relevant chunks on series access. ... This adds interval metadata to indexed chunks. The queried interval is used to filter chunks when queried from the index to save unnecessary accesses of the chunks file. This is especially relevant for series that come and go often and larger files.	8 years ago
Fabian Reinartz	bd77103a49	Add stats serialization, load querier of all blocks	8 years ago
Fabian Reinartz	d56b281006	Rename Iterator to Postings	8 years ago
Fabian Reinartz	ca89080128	Misc fixes for initial Prometheus integration	8 years ago
Fabian Reinartz	14dbc59f2b	cleanup and switching removal of unsafe calls.	8 years ago
Fabian Reinartz	40a451694f	Refactor persistence into interfaces	8 years ago
Fabian Reinartz	74f8dfd95d	Persist blocks periodically	8 years ago
Fabian Reinartz	0cf8bb9e53	Move sub-indexes into single index structure	8 years ago
Fabian Reinartz	8aa99a3ebd	misc	8 years ago
Fabian Reinartz	2c34a15fe6	Add initial seriailization of block data	8 years ago
Fabian Reinartz	3ef7da33c8	Restructure files	8 years ago