YDB storage plugin for Jaeger

This is a storage backend using Yandex.Database for Jaeger

components

• collector/query: these are docker images with jaeger-collector/jaeger-query and plugin built-in into image
• watcher (see cmd/schema/watcher): this is schema-watcher that creates new tables for spans/indexes and removes old ones

docker images

• cr.yandex/yc/jaeger-ydb-collector
• cr.yandex/yc/jaeger-ydb-query
• cr.yandex/yc/jaeger-ydb-watcher

how to run

cp docker-compose.example.yml docker-compose.yml
# edit docker-compose yml to set db dsn
docker-compose up -d

environment variables

Name	Type	Default	Description
YDB_ADDRESS	string		db endpoint host:port to connect to
YDB_SA_META_AUTH	bool	false	use metadata to authorize requests (documentation)
YDB_SA_ID	string		service account id for Yandex.Cloud authorization (doc on service accounts: https://cloud.yandex.com/docs/iam/concepts/users/service-accounts)
YDB_SA_KEY_ID	string		service account key id for Yandex.Cloud authorization
YDB_SA_PRIVATE_KEY_FILE	string		path to service account private key for Yandex.Cloud authorization
YDB_PATH	string		database path
YDB_FOLDER	string		folder to store data in)
YDB_CONNECT_TIMEOUT	duration	10s	db connect timeout
YDB_WRITE_TIMEOUT	duration	1s	write queries timeout
YDB_READ_TIMEOUT	duration	10s	read queries timeout
YDB_READ_QUERY_PARALLEL	integer	16	controls number of parallel read subqueries
YDB_READ_OP_LIMIT	integer	5000	max operation names to fetch for service
YDB_READ_SVC_LIMIT	integer	1000	max service names to fetch
YDB_POOL_SIZE	integer	100	db session pool size
YDB_QUERY_CACHE_SIZE	integer	50	db query cache size
YDB_WRITER_BUFFER_SIZE	integer	1000	span buffer size for batch writer
YDB_WRITER_BATCH_SIZE	integer	100	number of spans in batch write calls
YDB_WRITER_BATCH_WORKERS	integer	10	number of workers processing batch writes
YDB_INDEXER_BUFFER_SIZE	integer	1000	span buffer size for indexer
YDB_INDEXER_MAX_TRACES	integer	100	maximum trace_id count in a sinigle index record
YDB_INDEXER_MAX_TTL	duration	5s	maximum amount of time for indexer to batch trace_idы for index records
YDB_SCHEMA_NUM_PARTITIONS	integer	10	number of partitioned tables per day. Changing it requires recreating full data set
YDB_TOKEN	string		auth token for internal purposes

Configuration options can be passed via config file. Use --grpc-storage-plugin.configuration-file to pass configuration to YDB Plugin. In case of watcher use --config for the same purpose.

schema watcher configuration

Name	Type	Default	Description
WATCHER_AGE	duration	24h	delete partition tables older than this value
WATCHER_INTERVAL	duration	5m	check interval
YDB_FEATURE_SPLIT_BY_LOAD	bool	false	enable table split by load feature
YDB_FEATURE_COMPRESSION	bool	false	enable table compression feature, used for span storage

conference talks

• https://youtu.be/nyt_e4ULrUo?t=660
• https://youtu.be/hXH_tRBxFnA?t=11471