prometheus/head.go

292 lines
6.5 KiB
Go
Raw Normal View History

2016-12-04 04:16:11 -08:00
package tsdb
import (
"math"
2016-12-19 13:37:03 -08:00
"os"
"sort"
2016-12-04 04:16:11 -08:00
"sync"
"github.com/fabxc/tsdb/chunks"
2016-12-21 00:39:01 -08:00
"github.com/fabxc/tsdb/labels"
2016-12-31 00:48:49 -08:00
"github.com/prometheus/client_golang/prometheus"
2016-12-04 04:16:11 -08:00
)
// HeadBlock handles reads and writes of time series data within a time window.
type HeadBlock struct {
2016-12-21 16:12:28 -08:00
mtx sync.RWMutex
// descs holds all chunk descs for the head block. Each chunk implicitly
// is assigned the index as its ID.
descs []*chunkDesc
// hashes contains a collision map of label set hashes of chunks
// to their position in the chunk desc slice.
hashes map[uint64][]int
values map[string]stringset // label names to possible values
postings *memPostings // postings lists for terms
2016-12-22 03:05:24 -08:00
wal *WAL
stats BlockStats
2016-12-04 04:16:11 -08:00
}
2016-12-22 11:00:24 -08:00
// OpenHeadBlock creates a new empty head block.
func OpenHeadBlock(dir string, baseTime int64) (*HeadBlock, error) {
2016-12-22 06:18:33 -08:00
wal, err := OpenWAL(dir)
2016-12-22 03:05:24 -08:00
if err != nil {
return nil, err
}
b := &HeadBlock{
2016-12-21 16:12:28 -08:00
descs: []*chunkDesc{},
hashes: map[uint64][]int{},
values: map[string]stringset{},
postings: &memPostings{m: make(map[term][]uint32)},
2016-12-22 03:05:24 -08:00
wal: wal,
}
b.stats.MinTime = baseTime
2016-12-22 06:18:33 -08:00
err = wal.ReadAll(&walHandler{
series: func(lset labels.Labels) {
b.create(lset.Hash(), lset)
},
sample: func(s hashedSample) {
if err := b.descs[s.ref].append(s.t, s.v); err != nil {
panic(err) // TODO(fabxc): cannot actually error
}
b.stats.SampleCount++
},
})
if err != nil {
return nil, err
}
2016-12-22 03:05:24 -08:00
return b, nil
}
// Close syncs all data and closes underlying resources of the head block.
func (h *HeadBlock) Close() error {
return h.wal.Close()
}
// Querier returns a new querier over the head block.
func (h *HeadBlock) Querier(mint, maxt int64) Querier {
return newBlockQuerier(h, h, mint, maxt)
}
// Chunk returns the chunk for the reference number.
func (h *HeadBlock) Chunk(ref uint32) (chunks.Chunk, error) {
2016-12-21 16:12:28 -08:00
if int(ref) >= len(h.descs) {
return nil, errNotFound
}
2016-12-21 16:12:28 -08:00
return h.descs[int(ref)].chunk, nil
}
func (h *HeadBlock) interval() (int64, int64) {
return h.stats.MinTime, h.stats.MaxTime
}
// Stats returns statisitics about the indexed data.
func (h *HeadBlock) Stats() (BlockStats, error) {
return h.stats, nil
}
// LabelValues returns the possible label values
func (h *HeadBlock) LabelValues(names ...string) (StringTuples, error) {
if len(names) != 1 {
return nil, errInvalidSize
}
var sl []string
2016-12-21 16:12:28 -08:00
for s := range h.values[names[0]] {
sl = append(sl, s)
}
sort.Strings(sl)
t := &stringTuples{
l: len(names),
s: sl,
}
return t, nil
}
// Postings returns the postings list iterator for the label pair.
2016-12-14 12:58:29 -08:00
func (h *HeadBlock) Postings(name, value string) (Postings, error) {
2016-12-21 16:12:28 -08:00
return h.postings.get(term{name: name, value: value}), nil
}
// Series returns the series for the given reference.
func (h *HeadBlock) Series(ref uint32, mint, maxt int64) (Series, error) {
2016-12-21 16:12:28 -08:00
if int(ref) >= len(h.descs) {
return nil, errNotFound
}
2016-12-21 16:12:28 -08:00
cd := h.descs[ref]
if !intervalOverlap(cd.firsTimestamp, cd.lastTimestamp, mint, maxt) {
return nil, nil
}
s := &chunkSeries{
labels: cd.lset,
chunks: []ChunkMeta{
{MinTime: h.stats.MinTime, Ref: 0},
},
chunk: func(ref uint32) (chunks.Chunk, error) {
return cd.chunk, nil
},
}
return s, nil
}
2016-12-04 04:16:11 -08:00
// get retrieves the chunk with the hash and label set and creates
// a new one if it doesn't exist yet.
2016-12-22 03:05:24 -08:00
func (h *HeadBlock) get(hash uint64, lset labels.Labels) (*chunkDesc, uint32) {
2016-12-21 16:12:28 -08:00
refs := h.hashes[hash]
for _, ref := range refs {
if cd := h.descs[ref]; cd.lset.Equals(lset) {
2016-12-22 03:05:24 -08:00
return cd, uint32(ref)
2016-12-04 04:16:11 -08:00
}
}
2016-12-22 03:05:24 -08:00
return nil, 0
}
func (h *HeadBlock) create(hash uint64, lset labels.Labels) *chunkDesc {
2016-12-04 04:16:11 -08:00
cd := &chunkDesc{
lset: lset,
chunk: chunks.NewXORChunk(int(math.MaxInt64)),
}
2016-12-21 16:12:28 -08:00
// Index the new chunk.
ref := len(h.descs)
h.descs = append(h.descs, cd)
2016-12-22 03:05:24 -08:00
h.hashes[hash] = append(h.hashes[hash], ref)
2016-12-21 16:12:28 -08:00
// Add each label pair as a term to the inverted index.
terms := make([]term, 0, len(lset))
for _, l := range lset {
terms = append(terms, term{name: l.Name, value: l.Value})
valset, ok := h.values[l.Name]
if !ok {
valset = stringset{}
h.values[l.Name] = valset
}
valset.set(l.Value)
}
h.postings.add(uint32(ref), terms...)
2016-12-04 04:16:11 -08:00
// For the head block there's exactly one chunk per series.
h.stats.ChunkCount++
h.stats.SeriesCount++
2016-12-09 01:00:14 -08:00
return cd
2016-12-04 04:16:11 -08:00
}
2016-12-31 00:48:49 -08:00
func (h *HeadBlock) appendBatch(samples []hashedSample, appended prometheus.Counter) error {
2016-12-22 03:05:24 -08:00
// Find head chunks for all samples and allocate new IDs/refs for
// ones we haven't seen before.
var (
2016-12-26 07:55:32 -08:00
newSeries []labels.Labels
newHashes []uint64
uniqueHashes = map[uint64]uint32{}
2016-12-22 03:05:24 -08:00
)
2016-12-21 16:12:28 -08:00
2016-12-22 11:57:00 -08:00
for i := range samples {
s := &samples[i]
2016-12-22 03:05:24 -08:00
cd, ref := h.get(s.hash, s.labels)
if cd != nil {
// TODO(fabxc): sample refs are only scoped within a block for
// now and we ignore any previously set value
s.ref = ref
continue
}
2016-12-26 07:55:32 -08:00
// There may be several samples for a new series in a batch.
// We don't want to reserve a new space for each.
if ref, ok := uniqueHashes[s.hash]; ok {
s.ref = ref
continue
}
2016-12-22 03:05:24 -08:00
s.ref = uint32(len(h.descs) + len(newSeries))
2016-12-26 07:55:32 -08:00
uniqueHashes[s.hash] = s.ref
2016-12-22 11:57:00 -08:00
2016-12-22 03:05:24 -08:00
newSeries = append(newSeries, s.labels)
newHashes = append(newHashes, s.hash)
}
2016-12-21 16:12:28 -08:00
2016-12-22 03:05:24 -08:00
// Write all new series and samples to the WAL and add it to the
// in-mem database on success.
if err := h.wal.Log(newSeries, samples); err != nil {
return err
}
2016-12-22 03:05:24 -08:00
for i, s := range newSeries {
h.create(newHashes[i], s)
}
2016-12-22 03:05:24 -08:00
var merr MultiError
for _, s := range samples {
// TODO(fabxc): ensure that this won't be able to actually error in practice.
if err := h.descs[s.ref].append(s.t, s.v); err != nil {
merr.Add(err)
continue
}
2016-12-31 00:48:49 -08:00
appended.Inc()
2016-12-22 03:05:24 -08:00
h.stats.SampleCount++
if s.t > h.stats.MaxTime {
h.stats.MaxTime = s.t
}
2016-12-09 01:00:14 -08:00
}
2016-12-22 03:05:24 -08:00
return merr.Err()
}
2016-12-19 13:37:03 -08:00
func (h *HeadBlock) persist(p string) (int64, error) {
sf, err := os.Create(chunksFileName(p))
if err != nil {
return 0, err
}
xf, err := os.Create(indexFileName(p))
if err != nil {
return 0, err
}
iw := newIndexWriter(xf)
sw := newSeriesWriter(sf, iw, h.stats.MinTime)
defer sw.Close()
defer iw.Close()
2016-12-21 16:12:28 -08:00
for ref, cd := range h.descs {
if err := sw.WriteSeries(uint32(ref), cd.lset, []*chunkDesc{cd}); err != nil {
2016-12-19 13:37:03 -08:00
return 0, err
}
}
if err := iw.WriteStats(h.stats); err != nil {
2016-12-19 13:37:03 -08:00
return 0, err
}
2016-12-21 16:12:28 -08:00
for n, v := range h.values {
s := make([]string, 0, len(v))
for x := range v {
s = append(s, x)
}
if err := iw.WriteLabelIndex([]string{n}, s); err != nil {
2016-12-19 13:37:03 -08:00
return 0, err
}
}
2016-12-21 16:12:28 -08:00
for t := range h.postings.m {
if err := iw.WritePostings(t.name, t.value, h.postings.get(t)); err != nil {
2016-12-19 13:37:03 -08:00
return 0, err
}
}
2016-12-19 13:37:03 -08:00
return iw.Size() + sw.Size(), nil
}