Implement a metrics Prometheus endpoint - #1242
Merged
Merged
Conversation
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Benchmark Index (community)
Details
| Benchmark suite | Current: 7e1de02 | Previous: 4b1d28e | Ratio |
|---|---|---|---|
Add one schema (0 existing) |
309 ms |
289 ms |
1.07 |
Add one schema (100 existing) |
37 ms |
45 ms |
0.82 |
Add one schema (1000 existing) |
95 ms |
103 ms |
0.92 |
Add one schema (10000 existing) |
1147 ms |
857 ms |
1.34 |
Update one schema (1 existing) |
27 ms |
29 ms |
0.93 |
Update one schema (101 existing) |
38 ms |
38 ms |
1 |
Update one schema (1001 existing) |
105 ms |
106 ms |
0.99 |
Update one schema (10001 existing) |
1015 ms |
867 ms |
1.17 |
Cached rebuild (1 existing) |
10 ms |
8 ms |
1.25 |
Cached rebuild (101 existing) |
12 ms |
10 ms |
1.20 |
Cached rebuild (1001 existing) |
46 ms |
28 ms |
1.64 |
Cached rebuild (10001 existing) |
352 ms |
206 ms |
1.71 |
Index 100 schemas |
642 ms |
605 ms |
1.06 |
Index 1000 schemas |
1447 ms |
1236 ms |
1.17 |
Index 10000 schemas |
12878 ms |
13239 ms |
0.97 |
Index 10000 schemas (custom meta-schema) |
15181 ms |
16038 ms |
0.95 |
Index 10000 schemas ($ref fan-out) |
15066 ms |
15520 ms |
0.97 |
This comment was automatically generated by workflow using github-action-benchmark.
Contributor
There was a problem hiding this comment.
Benchmark Index (enterprise)
Details
| Benchmark suite | Current: 7e1de02 | Previous: 4b1d28e | Ratio |
|---|---|---|---|
Add one schema (0 existing) |
378 ms |
376 ms |
1.01 |
Add one schema (100 existing) |
127 ms |
126 ms |
1.01 |
Add one schema (1000 existing) |
180 ms |
187 ms |
0.96 |
Add one schema (10000 existing) |
822 ms |
823 ms |
1.00 |
Update one schema (1 existing) |
113 ms |
117 ms |
0.97 |
Update one schema (101 existing) |
127 ms |
122 ms |
1.04 |
Update one schema (1001 existing) |
180 ms |
187 ms |
0.96 |
Update one schema (10001 existing) |
837 ms |
839 ms |
1.00 |
Cached rebuild (1 existing) |
12 ms |
12 ms |
1 |
Cached rebuild (101 existing) |
18 ms |
15 ms |
1.20 |
Cached rebuild (1001 existing) |
45 ms |
45 ms |
1 |
Cached rebuild (10001 existing) |
358 ms |
363 ms |
0.99 |
Index 100 schemas |
553 ms |
481 ms |
1.15 |
Index 1000 schemas |
1537 ms |
1526 ms |
1.01 |
Index 10000 schemas |
13042 ms |
12951 ms |
1.01 |
Index 10000 schemas (custom meta-schema) |
15204 ms |
15426 ms |
0.99 |
Index 10000 schemas ($ref fan-out) |
15642 ms |
15728 ms |
0.99 |
This comment was automatically generated by workflow using github-action-benchmark.
Signed-off-by: Juan Cruz Viotti <jv@jviotti.com>
jviotti
changed the title
jviotti
marked this pull request as ready for review
August 25, 2026 14:40
🤖 Augment PR SummarySummary: This PR introduces a Prometheus-compatible metrics endpoint for Sourcemeta One.
🤖 Was this summary useful? React with 👍 or 👎 |
![augmentcode[bot]](https://avatars.githubusercontent.com/in/1027498?s=60&v=4){kind=small}
| auto snapshot = std::make_shared<HTTPRequest>( | ||
| std::string{this->method()}, std::string{this->path()}, | ||
| this->response_encoding_, raw_response); | ||
| snapshot->observation_ = this->observation_; |
There was a problem hiding this comment.
src/http/include/sourcemeta/one/http_request.h:182: a request aborted while body() is waiting never calls Observation::record()—the onAborted handler only flips completed. Since construction has already called enter(), each such disconnect permanently raises sourcemeta_one_http_requests_in_flight, so its value becomes unreliable and can grow without bound under aborted uploads.
Severity: medium
🤖 Was this useful? React with 👍 or 👎, or 🚀 if it prevented an incident/outage.
![cubic-dev-ai[bot]](https://avatars.githubusercontent.com/in/1082092?s=60&v=4){kind=small}
There was a problem hiding this comment.
3 issues found across 22 files
Prompt for AI agents (unresolved issues)
Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="src/http/include/sourcemeta/one/http_request.h">
<violation number="1" location="src/http/include/sourcemeta/one/http_request.h:53">
P2: When a handler throws or an asynchronous body request is aborted, this increment is never paired with `Observation::record()`. The Prometheus `in_flight` value then remains permanently elevated and the request is missing from answered totals; centralize completion with an idempotent guard that records error and abort paths too.</violation>
</file>
<file name="src/http/include/sourcemeta/one/http_metrics.h">
<violation number="1" location="src/http/include/sourcemeta/one/http_metrics.h:64">
P3: The exported `process_start_time_seconds` is recorded when `Router` calls `start()`, after process initialization and route loading. Capture the timestamp at process/global initialization, or rename the metric to represent server start.</violation>
</file>
<file name="enterprise/e2e/auth/prometheus.yml">
<violation number="1" location="enterprise/e2e/auth/prometheus.yml:15">
P3: The metrics bearer credential is duplicated as a literal in three files (`environment`, `prometheus.yml`, and every `Authorization` header in `metrics.all.hurl`) instead of being sourced from the single `ONE_E2E_KEY_METRICS` definition. Rotating the secret requires editing every site, and a mismatch breaks the scraper or the tests without an obvious cause. Inject the value into the prometheus service (e.g. `env_file: environment` plus `credentials: ${ONE_E2E_KEY_METRICS}`) so it follows the env-var convention already used by the other keys.</violation>
</file>
Reply with feedback, questions, or to request a fix.
Re-trigger cubic
| : request_{request}, response_{response} {} | ||
| : request_{request}, response_{response} { | ||
| this->observation_.started = std::chrono::steady_clock::now(); | ||
| http_metrics().enter(); |
There was a problem hiding this comment.
P2: When a handler throws or an asynchronous body request is aborted, this increment is never paired with Observation::record(). The Prometheus in_flight value then remains permanently elevated and the request is missing from answered totals; centralize completion with an idempotent guard that records error and abort paths too.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At src/http/include/sourcemeta/one/http_request.h, line 53:
<comment>When a handler throws or an asynchronous body request is aborted, this increment is never paired with `Observation::record()`. The Prometheus `in_flight` value then remains permanently elevated and the request is missing from answered totals; centralize completion with an idempotent guard that records error and abort paths too.</comment>
<file context>
@@ -25,12 +28,30 @@ namespace sourcemeta::one {
- : request_{request}, response_{response} {}
+ : request_{request}, response_{response} {
+ this->observation_.started = std::chrono::steady_clock::now();
+ http_metrics().enter();
+ }
</file context>
| // counted before this is said | ||
| auto start(const std::size_t handlers) -> void { | ||
| this->handlers_ = handlers; | ||
| this->started_ = |
There was a problem hiding this comment.
P3: The exported process_start_time_seconds is recorded when Router calls start(), after process initialization and route loading. Capture the timestamp at process/global initialization, or rename the metric to represent server start.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At src/http/include/sourcemeta/one/http_metrics.h, line 64:
<comment>The exported `process_start_time_seconds` is recorded when `Router` calls `start()`, after process initialization and route loading. Capture the timestamp at process/global initialization, or rename the metric to represent server start.</comment>
<file context>
@@ -0,0 +1,206 @@
+ // counted before this is said
+ auto start(const std::size_t handlers) -> void {
+ this->handlers_ = handlers;
+ this->started_ =
+ std::chrono::duration<double>{
+ std::chrono::system_clock::now().time_since_epoch()}
</file context>
| # like any other caller | ||
| authorization: | ||
| type: Bearer | ||
| credentials: metrics-secret-key |
There was a problem hiding this comment.
P3: The metrics bearer credential is duplicated as a literal in three files (environment, prometheus.yml, and every Authorization header in metrics.all.hurl) instead of being sourced from the single ONE_E2E_KEY_METRICS definition. Rotating the secret requires editing every site, and a mismatch breaks the scraper or the tests without an obvious cause. Inject the value into the prometheus service (e.g. env_file: environment plus credentials: ${ONE_E2E_KEY_METRICS}) so it follows the env-var convention already used by the other keys.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At enterprise/e2e/auth/prometheus.yml, line 15:
<comment>The metrics bearer credential is duplicated as a literal in three files (`environment`, `prometheus.yml`, and every `Authorization` header in `metrics.all.hurl`) instead of being sourced from the single `ONE_E2E_KEY_METRICS` definition. Rotating the secret requires editing every site, and a mismatch breaks the scraper or the tests without an obvious cause. Inject the value into the prometheus service (e.g. `env_file: environment` plus `credentials: ${ONE_E2E_KEY_METRICS}`) so it follows the env-var convention already used by the other keys.</comment>
<file context>
@@ -0,0 +1,18 @@
+ # like any other caller
+ authorization:
+ type: Bearer
+ credentials: metrics-secret-key
+ static_configs:
+ - targets:
</file context>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Signed-off-by: Juan Cruz Viotti jv@jviotti.com