You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Tracing spans record source URLs with credentials (remote_storage download_web*) #2817
#[tracing::instrument] on RemoteStorage::download_web_placed and on RemoteStorage::download_web_with_verification_and_caching (crates/common/src/remote_storage.rs) records the url argument in the span. A source URL can carry user:password@ and a signed query token. Any subscriber that records span fields, such as an OTLP exporter or a JSON log, then stores the secret. The log lines in crates/op/src/sources.rs already strip URLs for this reason (strip_urls, stripped_url).
What to do
Add skip(url) to both attributes, and record a stripped URL as a span field instead (fields(url = %stripped_url(url))).
Check the other #[tracing::instrument] attributes in crates/common and crates/op that take a URL.
### D2817 — The download_web spans of RemoteStorage record the source URL with no credentials, no query and no fragment
kind: task
size: S
est: 25 code, 70 tests
executor: agent
origin: residual
epic: 1437
layers:
- spans:
- crates/common/src/remote_storage.rs: Add skip(url) to the instrument attributes of download_web_with_verification_and_caching and download_web_placed, and record fields(url) from the stripped URL. Add a pub stripped_url function and a test that reads the span fields.
- crates/common/src/lib.rs: Export stripped_url from remote_storage.
- crates/op/src/specs.rs: stripped_url calls common::stripped_url, so one function strips a URL. strip_urls and its callers in sources.rs stay as they are.
test: cargo test -p common download_web_spans_record_a_stripped_url
proof: A subscriber that records span fields shows https://host/path for a source URL that carries user:password@ and a signed query, and the password and the token are not in the span.
cargo test -p common download_web_spans_record_a_stripped_url
0
running 1 test test remote_storage::span_tests::download_web_spans_record_a_stripped_url ... ok test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 91 filtered out; finished in 0.12s running 0 tests test result: ok. 0 passed; 0 failed; 0 ignored; 0 measured; 1 filtered out; finished in
Verify min task run agent-verify: exit 100 — FAIL [ 90.434s] (159/159) common::shell_harnesses all_script_test_harnesses_pass thread 'all_script_test_harnesses_pass' (2526) panicked at crates/common/tests/shell_harnesses.rs:83:5: shell test harnesses failed: metrics_test.sh failed (exit status: 1): == snapshot_validate
That /implement is held, not dropped. The container platform had no instance to give the coder: its pool of prepared containers was full, which happens most just after a deploy. It is asked again every 10 minutes for up to 30.
Nothing to type again: the command gets a rocket when its coder starts, or a refusal that names the reason.
What the platform said
could not start the coder: Error: there is no container instance that can be provided to this durable object
That /implement is queued, not refused. 10 follow-up coders are in flight, and 10 is the most that run at once, so no coder started. 1 queued command is ahead of it. The runtime starts queued follow-up coders oldest first as room frees, looking every ten minutes, and refuses one still waiting after 14 days.
Nothing to type again: the command gets a rocket when its coder starts, or a refusal that names the reason. /implement stop takes it off the queue.
That /implement is held, not dropped. The container platform had no instance to give the coder: its pool of prepared containers was full, which happens most just after a deploy. It is asked again every 10 minutes for up to 30.
Nothing to type again: the command gets a rocket when its coder starts, or a refusal that names the reason.
What the platform said
could not start the coder: Error: there is no container instance that can be provided to this durable object
Activity
norrietaylor commentedon Nov 20, 2026
Task
Source: #2817
Plan block
norrietaylor commentedon Nov 20, 2026
/implement
minimal-foundry commentedon Nov 20, 2026
Pull request #3713 is open for D2817.
branch
agent/2817-the-download-web-spans-of-remotestorage-record-t· epic #1437cargo test -p common download_web_spans_record_a_stripped_urlVerify
min task run agent-verify: exit 100 — FAIL [ 90.434s] (159/159) common::shell_harnesses all_script_test_harnesses_pass thread 'all_script_test_harnesses_pass' (2526) panicked at crates/common/tests/shell_harnesses.rs:83:5: shell test harnesses failed: metrics_test.sh failed (exit status: 1): == snapshot_validateUpdated 2026-11-20T22:51:12.148Z.
minimal-foundry commentedon Nov 20, 2026
That
/implementis held, not dropped. The container platform had no instance to give the coder: its pool of prepared containers was full, which happens most just after a deploy. It is asked again every 10 minutes for up to 30.Nothing to type again: the command gets a rocket when its coder starts, or a refusal that names the reason.
What the platform said
could not start the coder: Error: there is no container instance that can be provided to this durable object
minimal-foundry commentedon Nov 20, 2026
That
/implementis queued, not refused. 10 follow-up coders are in flight, and 10 is the most that run at once, so no coder started. 1 queued command is ahead of it. The runtime starts queued follow-up coders oldest first as room frees, looking every ten minutes, and refuses one still waiting after 14 days.Nothing to type again: the command gets a rocket when its coder starts, or a refusal that names the reason.
/implement stoptakes it off the queue.minimal-foundry commentedon Nov 20, 2026
That
/implementis held, not dropped. The container platform had no instance to give the coder: its pool of prepared containers was full, which happens most just after a deploy. It is asked again every 10 minutes for up to 30.Nothing to type again: the command gets a rocket when its coder starts, or a refusal that names the reason.
What the platform said
could not start the coder: Error: there is no container instance that can be provided to this durable object