rust/greptimedb
1
0
Fork 0
mirror of https://github.com/GreptimeTeam/greptimedb.git synced 2026-01-06 21:32:58 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat: opentelemetry trace new data modeling (#5622)

Browse Source 
* feat: include trace v1 encoding

* feat: add trace ingestion in inserter

* feat: add partition rules and index for trace_id

* chore: format

* chore: fmt

* fix: issue introduced with merge

* feat: adjust index and add integration test for v1

* refactor: remove comment key

* fix: update default value of skip index granularity

* fix: update default value of skip index granularity

* refactor: rename some functions

* feat: remove skipping index from span_id

* refactor: made span_id part of primary key for potential dedup purpose

* feat: move the special attribute resource_attribute.service.name to top level

---------

Co-authored-by: shuiyisong <113876041+shuiyisong@users.noreply.github.com>
This commit is contained in:
Ning Sun
2025-03-04 20:08:52 -08:00
committed by GitHub
parent b90ef10523
commit 37f8341963
16 changed files with 714 additions and 97 deletions
Download Patch File Download Diff File Expand all files Collapse all files

100
tests-integration/tests/http.rs
View File

@@ -96,7 +96,8 @@ macro_rules! http_tests {
test_pipeline_dispatcher,
test_otlp_metrics,
test_otlp_traces,
test_otlp_traces_v0,
test_otlp_traces_v1,
test_otlp_logs,
test_loki_pb_logs,
test_loki_json_logs,
@@ -1320,7 +1321,7 @@ transform:
// 3. check schema
let expected_schema = "[[\"logs1\",\"CREATE TABLE IF NOT EXISTS \\\"logs1\\\" (\\n \\\"id1\\\" INT NULL,\\n \\\"id2\\\" INT NULL,\\n \\\"logger\\\" STRING NULL,\\n \\\"type\\\" STRING NULL SKIPPING INDEX WITH(granularity = '0', type = 'BLOOM'),\\n \\\"log\\\" STRING NULL FULLTEXT INDEX WITH(analyzer = 'English', case_sensitive = 'false'),\\n \\\"time\\\" TIMESTAMP(9) NOT NULL,\\n TIME INDEX (\\\"time\\\")\\n)\\n\\nENGINE=mito\\nWITH(\\n append_mode = 'true'\\n)\"]]";
let expected_schema = "[[\"logs1\",\"CREATE TABLE IF NOT EXISTS \\\"logs1\\\" (\\n \\\"id1\\\" INT NULL,\\n \\\"id2\\\" INT NULL,\\n \\\"logger\\\" STRING NULL,\\n \\\"type\\\" STRING NULL SKIPPING INDEX WITH(granularity = '10240', type = 'BLOOM'),\\n \\\"log\\\" STRING NULL FULLTEXT INDEX WITH(analyzer = 'English', case_sensitive = 'false'),\\n \\\"time\\\" TIMESTAMP(9) NOT NULL,\\n TIME INDEX (\\\"time\\\")\\n)\\n\\nENGINE=mito\\nWITH(\\n append_mode = 'true'\\n)\"]]";
validate_data(
"pipeline_schema",
&client,
@@ -2053,7 +2054,7 @@ pub async fn test_otlp_metrics(store_type: StorageType) {
guard.remove_all().await;
}
pub async fn test_otlp_traces(store_type: StorageType) {
pub async fn test_otlp_traces_v0(store_type: StorageType) {
// init
common_telemetry::init_default_ut_logging();
let (app, mut guard) = setup_test_http_app_with_frontend(store_type, "test_otlp_traces").await;
@@ -2125,6 +2126,99 @@ pub async fn test_otlp_traces(store_type: StorageType) {
guard.remove_all().await;
}
pub async fn test_otlp_traces_v1(store_type: StorageType) {
// init
common_telemetry::init_default_ut_logging();
let (app, mut guard) = setup_test_http_app_with_frontend(store_type, "test_otlp_traces").await;
const TRACE_V1: &str = "greptime_trace_v1";
let content = r#"
{"resourceSpans":[{"resource":{"attributes":[{"key":"service.name","value":{"stringValue":"telemetrygen"}}],"droppedAttributesCount":0},"scopeSpans":[{"scope":{"name":"telemetrygen","version":"","attributes":[],"droppedAttributesCount":0},"spans":[{"traceId":"c05d7a4ec8e1f231f02ed6e8da8655b4","spanId":"9630f2916e2f7909","traceState":"","parentSpanId":"d24f921c75f68e23","flags":256,"name":"okey-dokey-0","kind":2,"startTimeUnixNano":"1736480942444376000","endTimeUnixNano":"1736480942444499000","attributes":[{"key":"net.peer.ip","value":{"stringValue":"1.2.3.4"}},{"key":"peer.service","value":{"stringValue":"telemetrygen-client"}}],"droppedAttributesCount":0,"events":[],"droppedEventsCount":0,"links":[],"droppedLinksCount":0,"status":{"message":"","code":0}},{"traceId":"c05d7a4ec8e1f231f02ed6e8da8655b4","spanId":"d24f921c75f68e23","traceState":"","parentSpanId":"","flags":256,"name":"lets-go","kind":3,"startTimeUnixNano":"1736480942444376000","endTimeUnixNano":"1736480942444499000","attributes":[{"key":"net.peer.ip","value":{"stringValue":"1.2.3.4"}},{"key":"peer.service","value":{"stringValue":"telemetrygen-server"}}],"droppedAttributesCount":0,"events":[],"droppedEventsCount":0,"links":[],"droppedLinksCount":0,"status":{"message":"","code":0}},{"traceId":"cc9e0991a2e63d274984bd44ee669203","spanId":"8f847259b0f6e1ab","traceState":"","parentSpanId":"eba7be77e3558179","flags":256,"name":"okey-dokey-0","kind":2,"startTimeUnixNano":"1736480942444589000","endTimeUnixNano":"1736480942444712000","attributes":[{"key":"net.peer.ip","value":{"stringValue":"1.2.3.4"}},{"key":"peer.service","value":{"stringValue":"telemetrygen-client"}}],"droppedAttributesCount":0,"events":[],"droppedEventsCount":0,"links":[],"droppedLinksCount":0,"status":{"message":"","code":0}},{"traceId":"cc9e0991a2e63d274984bd44ee669203","spanId":"eba7be77e3558179","traceState":"","parentSpanId":"","flags":256,"name":"lets-go","kind":3,"startTimeUnixNano":"1736480942444589000","endTimeUnixNano":"1736480942444712000","attributes":[{"key":"net.peer.ip","value":{"stringValue":"1.2.3.4"}},{"key":"peer.service","value":{"stringValue":"telemetrygen-server"}}],"droppedAttributesCount":0,"events":[],"droppedEventsCount":0,"links":[],"droppedLinksCount":0,"status":{"message":"","code":0}}],"schemaUrl":""}],"schemaUrl":"https://opentelemetry.io/schemas/1.4.0"}]}
"#;
let req: ExportTraceServiceRequest = serde_json::from_str(content).unwrap();
let body = req.encode_to_vec();
// handshake
let client = TestClient::new(app).await;
// write traces data
let res = send_req(
&client,
vec![
(
HeaderName::from_static("content-type"),
HeaderValue::from_static("application/x-protobuf"),
),
(
HeaderName::from_static("x-greptime-log-pipeline-name"),
HeaderValue::from_static(TRACE_V1),
),
(
HeaderName::from_static("x-greptime-trace-table-name"),
HeaderValue::from_static("mytable"),
),
],
"/v1/otlp/v1/traces",
body.clone(),
false,
)
.await;
assert_eq!(StatusCode::OK, res.status());
// select traces data
let expected = r#"[[1736480942444376000,1736480942444499000,123000,"c05d7a4ec8e1f231f02ed6e8da8655b4","9630f2916e2f7909","d24f921c75f68e23","SPAN_KIND_SERVER","okey-dokey-0","STATUS_CODE_UNSET","","","telemetrygen","","telemetrygen","1.2.3.4","telemetrygen-client",[],[]],[1736480942444376000,1736480942444499000,123000,"c05d7a4ec8e1f231f02ed6e8da8655b4","d24f921c75f68e23","","SPAN_KIND_CLIENT","lets-go","STATUS_CODE_UNSET","","","telemetrygen","","telemetrygen","1.2.3.4","telemetrygen-server",[],[]],[1736480942444589000,1736480942444712000,123000,"cc9e0991a2e63d274984bd44ee669203","8f847259b0f6e1ab","eba7be77e3558179","SPAN_KIND_SERVER","okey-dokey-0","STATUS_CODE_UNSET","","","telemetrygen","","telemetrygen","1.2.3.4","telemetrygen-client",[],[]],[1736480942444589000,1736480942444712000,123000,"cc9e0991a2e63d274984bd44ee669203","eba7be77e3558179","","SPAN_KIND_CLIENT","lets-go","STATUS_CODE_UNSET","","","telemetrygen","","telemetrygen","1.2.3.4","telemetrygen-server",[],[]]]"#;
validate_data("otlp_traces", &client, "select * from mytable;", expected).await;
let expected_ddl = r#"[["mytable","CREATE TABLE IF NOT EXISTS \"mytable\" (\n \"timestamp\" TIMESTAMP(9) NOT NULL,\n \"timestamp_end\" TIMESTAMP(9) NULL,\n \"duration_nano\" BIGINT UNSIGNED NULL,\n \"trace_id\" STRING NULL SKIPPING INDEX WITH(granularity = '10240', type = 'BLOOM'),\n \"span_id\" STRING NULL,\n \"parent_span_id\" STRING NULL SKIPPING INDEX WITH(granularity = '10240', type = 'BLOOM'),\n \"span_kind\" STRING NULL,\n \"span_name\" STRING NULL SKIPPING INDEX WITH(granularity = '10240', type = 'BLOOM'),\n \"span_status_code\" STRING NULL,\n \"span_status_message\" STRING NULL,\n \"trace_state\" STRING NULL,\n \"scope_name\" STRING NULL,\n \"scope_version\" STRING NULL,\n \"service_name\" STRING NULL,\n \"span_attributes.net.peer.ip\" STRING NULL,\n \"span_attributes.peer.service\" STRING NULL,\n \"span_events\" JSON NULL,\n \"span_links\" JSON NULL,\n TIME INDEX (\"timestamp\"),\n PRIMARY KEY (\"trace_id\", \"span_id\")\n)\nPARTITION ON COLUMNS (\"trace_id\") (\n trace_id < '1',\n trace_id >= '1' AND trace_id < '2',\n trace_id >= '2' AND trace_id < '3',\n trace_id >= '3' AND trace_id < '4',\n trace_id >= '4' AND trace_id < '5',\n trace_id >= '5' AND trace_id < '6',\n trace_id >= '6' AND trace_id < '7',\n trace_id >= '7' AND trace_id < '8',\n trace_id >= '8' AND trace_id < '9',\n trace_id >= '9' AND trace_id < 'A',\n trace_id >= 'A' AND trace_id < 'B' OR trace_id >= 'a' AND trace_id < 'b',\n trace_id >= 'B' AND trace_id < 'C' OR trace_id >= 'b' AND trace_id < 'c',\n trace_id >= 'C' AND trace_id < 'D' OR trace_id >= 'c' AND trace_id < 'd',\n trace_id >= 'D' AND trace_id < 'E' OR trace_id >= 'd' AND trace_id < 'e',\n trace_id >= 'E' AND trace_id < 'F' OR trace_id >= 'e' AND trace_id < 'f',\n trace_id >= 'F' AND trace_id < 'a' OR trace_id >= 'f'\n)\nENGINE=mito\nWITH(\n append_mode = 'true'\n)"]]"#;
validate_data(
"otlp_traces",
&client,
"show create table mytable;",
expected_ddl,
)
.await;
// drop table
let res = client.get("/v1/sql?sql=drop table mytable;").send().await;
assert_eq!(res.status(), StatusCode::OK);
// write traces data with gzip
let res = send_req(
&client,
vec![
(
HeaderName::from_static("content-type"),
HeaderValue::from_static("application/x-protobuf"),
),
(
HeaderName::from_static("x-greptime-log-pipeline-name"),
HeaderValue::from_static(TRACE_V1),
),
(
HeaderName::from_static("x-greptime-trace-table-name"),
HeaderValue::from_static("mytable"),
),
],
"/v1/otlp/v1/traces",
body.clone(),
true,
)
.await;
assert_eq!(StatusCode::OK, res.status());
// select traces data again
validate_data(
"otlp_traces_with_gzip",
&client,
"select * from mytable;",
expected,
)
.await;
guard.remove_all().await;
}
pub async fn test_otlp_logs(store_type: StorageType) {
common_telemetry::init_default_ut_logging();
let (app, mut guard) = setup_test_http_app_with_frontend(store_type, "test_otlp_logs").await;