fix: deflake //rs/tests/consensus/tecdsa:tecdsa_add_nodes_test

[basvandijk]

The //rs/tests/consensus/tecdsa:tecdsa_add_nodes_test is slightly flaky:

$ bazel run //ci/githubstats:query -- top 1 flaky% --week --include //rs/tests/consensus/tecdsa:tecdsa_add_nodes_test
...
┍━━━━┯━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┯━━━━━━━━━┯━━━━━━━━━━━━━━━┯━━━━━━━━━┯━━━━━━━━━━━┯━━━━━━━━┯━━━━━━━━━━━━━━━━┯━━━━━━━━━━┯━━━━━━━━━━━━┯━━━━━━━━━┯━━━━━━━━━━┯━━━━━━━━━━━━━━━━━━┯━━━━━━━━━━━━━━━━┯━━━━━━━━━━━┑
│    │ label                                             │   total │   non_success │   flaky │   timeout │   fail │   non_success% │   flaky% │   timeout% │   fail% │   impact │   total duration │   duration_p90 │ owners    │
┝━━━━┿━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┿━━━━━━━━━┿━━━━━━━━━━━━━━━┿━━━━━━━━━┿━━━━━━━━━━━┿━━━━━━━━┿━━━━━━━━━━━━━━━━┿━━━━━━━━━━┿━━━━━━━━━━━━┿━━━━━━━━━┿━━━━━━━━━━┿━━━━━━━━━━━━━━━━━━┿━━━━━━━━━━━━━━━━┿━━━━━━━━━━━┥
│  0 │ //rs/tests/consensus/tecdsa:tecdsa_add_nodes_test │     121 │             3 │       3 │         0 │      0 │            2.5 │      2.5 │          0 │       0 │    18:30 │         12:26:10 │           6:10 │ consensus │
┕━━━━┷━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┷━━━━━━━━━┷━━━━━━━━━━━━━━━┷━━━━━━━━━┷━━━━━━━━━━━┷━━━━━━━━┷━━━━━━━━━━━━━━━━┷━━━━━━━━━━┷━━━━━━━━━━━━┷━━━━━━━━━┷━━━━━━━━━━┷━━━━━━━━━━━━━━━━━━┷━━━━━━━━━━━━━━━━┷━━━━━━━━━━━┙

Failing for different reasons:

$ bazel run //ci/githubstats:query -- last --flaky --week //rs/tests/consensus/tecdsa:tecdsa_add_nodes_test  --columns=last_started_at,errors
...
╒════╤═════════════════════════╤════════════════════════════════════════════════════════════════════════╕
│    │   last started at (UTC) │ errors per attempt                                                     │
╞════╪═════════════════════════╪════════════════════════════════════════════════════════════════════════╡
│  0 │ Fri 2026-02-20 11:36:51 │ 1: test: Timeout after 600s                                            │
├────┼─────────────────────────┼────────────────────────────────────────────────────────────────────────┤
│  1 │ Thu 2026-02-19 15:32:17 │ 1: test: Failed to find key rotation of ecdsa:Secp256k1:some_ecdsa_key │
├────┼─────────────────────────┼────────────────────────────────────────────────────────────────────────┤
│  2 │ Mon 2026-02-16 11:39:59 │ 1: test: nns canisters were not installed within timeout of 160 sec    │
╘════╧═════════════════════════╧════════════════════════════════════════════════════════════════════════╛

Claude Opus 4.6 made the following Root Cause Analysis and accompanying fixes:

Root Cause

The test's per-task timeout of 600s (default DEFAULT_TIMEOUT_PER_TEST) is too tight for the operations performed in the test function. The test does NNS canister installation, chain key operations, node addition, topology propagation, key rotation verification, progress checks, and signature tests — all within a single test function that gets 600s.

Three failure modes were observed across 3 flaky runs in the past week:

1. NNS canister install timeout — NNS install within the test function consumed most of the time budget (Feb 16)
2. Key rotation assertion failure — assert_metric_sum failed after only 30 retries × 3s = 90s, which was not enough time for key rotation to complete after adding 3 nodes (Feb 19)
3. Overall 600s per-task timeout — Test hit the 600s timeout while waiting for topology propagation via await_subnet_earliest_topology_version (Feb 20)

Changes

• The NNS_CANISTER_INSTALL_TIMEOUT was already increased from 160s to 300s by:
  fix: deflake //rs/tests/nns:bitcoin_set_config_proposal_test #8951.
• Set overall_timeout and timeout_per_test to 900s (matching Bazel's long test timeout), up from the 600s default
• Increase assert_metric_sum retries from 30 to 60 (180s max wait for key rotation, up from 90s)

Skill: .claude/skills/fix-flaky-tests/SKILL.md.