Skip to content

Record metrics and charge for HTTP route egress#5505

Open
bradleyshep wants to merge 11 commits into
masterfrom
bradley/record-metrics-and-charge-http
Open

Record metrics and charge for HTTP route egress#5505
bradleyshep wants to merge 11 commits into
masterfrom
bradley/record-metrics-and-charge-http

Conversation

@bradleyshep

Copy link
Copy Markdown
Contributor

Description of Changes

The HTTP /sql endpoint charged for its DB work (rows/bytes scanned, writes) but not for egress, the bytes sent back to the client. WebSocket queries already count these, but the HTTP path skipped it, so the same query was billed over WS and free over HTTP. There were also no metrics on the HTTP routes.

This PR:

  1. Charges egress for /sql, using the BSATN size of the result rows so the price matches WebSocket queries.
  2. Adds a /v1 middleware that records per-route request count, latency, and body sizes. Body sizes count bytes actually transferred through a small Body wrapper rather than trusting the Content-Length header, so chunked and streamed bodies are counted and spoofed headers are not. Non-standard HTTP methods are bucketed into an OTHER label to keep cardinality bounded.

Adds http-body as a direct dependency (already in the tree via axum/hyper).

Not covered here: CORS preflights and unmatched 404s are not counted (they short-circuit before the middleware), and routes other than /sql are not billed egress yet.

API and ABI breaking changes

None. The /sql response is unchanged, only internal metrics were added.

Expected complexity level and risk

  1. The egress change is one counter increment on the existing SQL path. The middleware is router setup plus a ~35 line body wrapper.

Testing

  • Unit test that a SELECT charges bytes_sent_to_clients equal to the rows' BSATN size.
  • Verified locally: a SELECT over HTTP bumps the egress counter by the exact BSATN size and the spacetime_http_* counters by the exact wire bytes (checked against curl's size_download).
  • Verified locally: a bogus method (curl -X BANANA) lands in the OTHER bucket, and a chunked request body (no Content-Length) is counted.
  • Existing SQL smoketests pass.

Comment thread crates/core/src/sql/execute.rs Outdated
Comment on lines +133 to +134
// Charge egress, using BSATN size for parity with WebSocket queries.
metrics.bytes_sent_to_clients += rows.iter().map(|row| bsatn::to_len(row).unwrap_or(0)).sum::<usize>();

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

This seems potentially redundant in a way that worries me. If we charge for both bytes_sent_to_clients and http_response_body_bytes, then customers get charged for egress due to SQL queries over HTTP twice. If we call run_inner from the WS path (which I don't think we do, at least currently, but I could easily imagine us doing in the future), then we'll double count on bytes_sent_to_clients directly.

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

Good to know. I will look into better future-proofing this, if that works.

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

@gefjon Thanks reviewing + catching that. Moved it into exec_sql, so run_inner doesn't touch egress anymore and each transport charges for what it sends (like WS already does...). Also added a test to make sure the executor never charges. The http_* metrics aren't billed, they're just for observability. Added a comment for clarity.

Each transport charges for the rows it sends, matching how WebSocket
queries charge at their serialization point. Prevents double counting
if the executor gains other transport callers.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants