Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

non-ambiguous internal aggregations #5715

Open
wants to merge 5 commits into
base: main
Choose a base branch
from
Open

non-ambiguous internal aggregations #5715

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
791efcd
add proxy struct for aggregations
trinity-1686a Feb 28, 2025
26f79c2
use postcard for aggregations internally
trinity-1686a Mar 4, 2025
e785fe1
add license header
trinity-1686a Mar 17, 2025
e9cab79
also fork percentile
trinity-1686a Mar 17, 2025
5b23c88
use vec instead of hashmaps
trinity-1686a Mar 31, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 4 additions & 2 deletions quickwit/Cargo.lock
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

1 change: 1 addition & 0 deletions quickwit/Cargo.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -205,6 +205,7 @@ reqwest = { version = "0.11", default-features = false, features = [
"rustls-tls",
] }
rust-embed = "6.8.1"
rustc-hash = "2.1.1"
rustls = "0.21"
rustls-pemfile = "1.0.0"
sea-query = { version = "0.30" }
Expand Down
2 changes: 1 addition & 1 deletion quickwit/quickwit-common/Cargo.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -39,7 +39,7 @@ thiserror = { workspace = true }
tokio = { workspace = true }
tokio-metrics = { workspace = true }
tokio-stream = { workspace = true }
tonic = { workspace = true }
tonic = { workspace = true, features = ["tls"] }
tower = { workspace = true }
tracing = { workspace = true }

Expand Down
1 change: 1 addition & 0 deletions quickwit/quickwit-jaeger/Cargo.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,7 @@ license.workspace = true
async-trait = { workspace = true }
itertools = { workspace = true }
once_cell = { workspace = true }
postcard = { workspace = true }
prost = { workspace = true }
prost-types = { workspace = true }
serde = { workspace = true }
Expand Down
77 changes: 54 additions & 23 deletions quickwit/quickwit-jaeger/src/lib.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -282,11 +282,11 @@ impl JaegerService {
};
let search_response = self.search_service.root_search(search_request).await?;

let Some(agg_result_json) = search_response.aggregation else {
let Some(agg_result_postcard) = search_response.aggregation_postcard else {
debug!("the query matched no traces");
return Ok((Vec::new(), 0..=0));
};
let trace_ids = collect_trace_ids(&agg_result_json)?;
let trace_ids = collect_trace_ids(&agg_result_postcard)?;
debug!("the query matched {} traces.", trace_ids.0.len());
Ok(trace_ids)
}
Expand Down Expand Up @@ -1087,9 +1087,11 @@ fn qw_event_to_jaeger_log(event: QwEvent) -> Result<JaegerLog, Status> {
Ok(log)
}

fn collect_trace_ids(trace_ids_json: &str) -> Result<(Vec<TraceId>, TimeIntervalSecs), Status> {
fn collect_trace_ids(
trace_ids_postcard: &[u8],
) -> Result<(Vec<TraceId>, TimeIntervalSecs), Status> {
let collector_fruit: <FindTraceIdsCollector as Collector>::Fruit =
json_deserialize(trace_ids_json, "trace IDs aggregation")?;
postcard_deserialize(trace_ids_postcard, "trace IDs aggregation")?;
if collector_fruit.is_empty() {
return Ok((Vec::new(), 0..=0));
}
Expand Down Expand Up @@ -1118,6 +1120,19 @@ where T: Deserialize<'a> {
}
}

fn postcard_deserialize<'a, T>(json: &'a [u8], label: &'static str) -> Result<T, Status>
where T: Deserialize<'a> {
match postcard::from_bytes(json) {
Ok(deserialized) => Ok(deserialized),
Err(error) => {
error!("failed to deserialize {label}: {error:?}");
Err(Status::internal(format!(
"Failed to deserialize {label}: {error:?}."
)))
}
}
}

#[cfg(test)]
mod tests {
use quickwit_opentelemetry::otlp::{OtelSignal, OTEL_TRACES_INDEX_ID_PATTERN};
Expand Down Expand Up @@ -2470,34 +2485,50 @@ mod tests {

#[test]
fn test_collect_trace_ids() {
use quickwit_opentelemetry::otlp::TraceId;
use quickwit_search::Span;
use tantivy::DateTime;
{
let agg_result_json = r#"[]"#;
let (trace_ids, _span_timestamps_range) = collect_trace_ids(agg_result_json).unwrap();
let agg_result: Vec<Span> = Vec::new();
let agg_result_postcard = postcard::to_stdvec(&agg_result).unwrap();
let (trace_ids, _span_timestamps_range) =
collect_trace_ids(&agg_result_postcard).unwrap();
assert!(trace_ids.is_empty());
}
{
let agg_result_json = r#"[
{
"trace_id": "01010101010101010101010101010101",
"span_timestamp": 1684857492783747000
}
]"#;
let (trace_ids, span_timestamps_range) = collect_trace_ids(agg_result_json).unwrap();
let agg_result = vec![Span {
trace_id: TraceId::new([
0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
0x01, 0x01, 0x01,
]),
span_timestamp: DateTime::from_timestamp_nanos(1684857492783747000),
}];
let agg_result_postcard = postcard::to_stdvec(&agg_result).unwrap();
let (trace_ids, span_timestamps_range) =
collect_trace_ids(&agg_result_postcard).unwrap();
assert_eq!(trace_ids.len(), 1);
assert_eq!(span_timestamps_range, 1684857492..=1684857492);
}
{
let agg_result_json = r#"[
{
"trace_id": "0102030405060708090a0b0c0d0e0f10",
"span_timestamp": 1684857492783747000
let agg_result = vec![
Span {
trace_id: TraceId::new([
0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b, 0x0c,
0x0d, 0x0e, 0x0f, 0x10,
]),
span_timestamp: DateTime::from_timestamp_nanos(1684857492783747000),
},
{
"trace_id": "02020202020202020202020202020202",
"span_timestamp": 1684857826019627000
}
]"#;
let (trace_ids, span_timestamps_range) = collect_trace_ids(agg_result_json).unwrap();
Span {
trace_id: TraceId::new([
0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,
0x02, 0x02, 0x02, 0x02,
]),
span_timestamp: DateTime::from_timestamp_nanos(1684857826019627000),
},
];
let agg_result_postcard = postcard::to_stdvec(&agg_result).unwrap();
let (trace_ids, span_timestamps_range) =
collect_trace_ids(&agg_result_postcard).unwrap();
assert_eq!(trace_ids.len(), 2);
assert_eq!(span_timestamps_range, 1684857492..=1684857826);
}
Expand Down
7 changes: 5 additions & 2 deletions quickwit/quickwit-proto/protos/quickwit/search.proto
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -289,8 +289,11 @@ message SearchResponse {
// The searcherrors that occurred formatted as string.
repeated string errors = 4;

// Serialized aggregation response
optional string aggregation = 5;
// used to be json-encoded aggregation
reserved 5;

// Postcard-encoded aggregation response
optional bytes aggregation_postcard = 9;
Comment on lines -292 to +296
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

this is not retro-compatible, right? it means we cannot perform a rolling upgrade of searchers with this change

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

you need the node which serves the rest api and the root-searcher for the request to be the same, otherwise the aggregation result just disappear, but you can have root-searcher and leaf-searcher on different versions and that's alright

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

you need the node which serves the rest api and the root-searcher for the request to be the same

From what I read in the code, this is always the case: the REST API uses SearchService that only has one concrete implementation (SearchServiceImpl).

Now, this raises a new question. We could also solve this by getting rid of the serialization entirely. If the SearchService trait was never re-implemented so far, we can assume it was a premature extra abstraction layer.

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

let's keep that abstraction


// Scroll Id (only set if scroll_secs was set in the request)
optional string scroll_id = 6;
Expand Down
Binary file added quickwit/quickwit-proto/src/codegen/cloudprem/descriptor.bin
View file
Open in desktop
Binary file not shown.
6 changes: 3 additions & 3 deletions quickwit/quickwit-proto/src/codegen/quickwit/quickwit.search.rs
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

1 change: 1 addition & 0 deletions quickwit/quickwit-query/Cargo.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,6 +26,7 @@ tantivy = { workspace = true }
tantivy-fst = { workspace = true }
time = { workspace = true }
thiserror = { workspace = true }
rustc-hash = { workspace = true }
whichlang = { workspace = true, optional = true }

quickwit-common = { workspace = true }
Expand Down
Loading
Loading