fix(forecast): surface silent forecast input feed gaps#5109
Merged
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub. 1 Skipped Deployment
|
koala73
force-pushed
the
codex/issue-5101-signal-observability
branch
from
July 9, 2026 14:02
843dd6f to
9642fdc
Compare
Contributor
Greptile SummaryThis PR surfaces silent forecast input feed failures by adding per-feed record-count logging, wiring temporal anomalies into the strict health-check path, and keeping the feed warm via relay and
Confidence Score: 4/5Safe to merge; the observability and health wiring is well-structured, tested, and consistent with existing patterns in the codebase. The core logic is sound — scripts/seed-forecasts.mjs — specifically the interaction between Important Files Changed
Sequence Diagram%%{init: {'theme': 'neutral'}}%%
sequenceDiagram
participant SI as seed-infra (warm-ping)
participant Relay as relay warm-ping
participant LTA as list-temporal-anomalies
participant Redis as Redis
participant SF as seed-forecasts (readInputKeys)
participant HC as /api/health
SI->>LTA: POST /api/infrastructure/v1/list-temporal-anomalies
Relay->>LTA: POST /api/infrastructure/v1/list-temporal-anomalies
LTA->>Redis: GET temporal:anomalies:v1
alt "cache hit (age < 900s)"
Redis-->>LTA: cached snapshot
LTA-->>SI: return cached (no seed-meta update)
else cache miss or stale
LTA->>LTA: compute fresh snapshot
LTA->>Redis: "SET temporal:anomalies:v1 (TTL=900s)"
Redis-->>LTA: "published=true"
LTA->>Redis: "SET seed-meta:temporal:anomalies {fetchedAt, recordCount} (TTL=604800s)"
LTA-->>SI: return snapshot
end
SF->>Redis: pipeline GET all forecast input keys
Redis-->>SF: parsedByKey
SF->>SF: buildForecastInputPresenceRows(parsedByKey)
SF->>SF: warnOnMissingForecastInputs(rows)
Note over SF: logs [ForecastInputs] warn per zero-record feed
HC->>Redis: GET temporal:anomalies:v1 strlen
HC->>Redis: GET seed-meta:temporal:anomalies
alt "data present AND seed-meta.fetchedAt < 45min ago"
HC-->>HC: "status=OK"
else missing or stale
HC-->>HC: "status=EMPTY or STALE_SEED (CRIT)"
end
%%{init: {'theme': 'base', 'themeVariables': {"darkMode": true, "background": "#0d1117", "primaryColor": "#21262d", "primaryTextColor": "#e6edf3", "primaryBorderColor": "#8b949e", "lineColor": "#8b949e", "textColor": "#e6edf3", "edgeLabelBackground": "#161b22", "actorBkg": "#21262d", "actorBorder": "#8b949e", "actorTextColor": "#e6edf3", "actorLineColor": "#8b949e", "signalColor": "#8b949e", "signalTextColor": "#e6edf3", "noteBkgColor": "#373320", "noteBorderColor": "#d4a72c", "noteTextColor": "#f0e6c0", "labelBoxBkgColor": "#21262d", "labelBoxBorderColor": "#8b949e", "labelTextColor": "#e6edf3", "loopTextColor": "#e6edf3", "activationBkgColor": "#30363d", "activationBorderColor": "#8b949e"}}}%%
sequenceDiagram
participant SI as seed-infra (warm-ping)
participant Relay as relay warm-ping
participant LTA as list-temporal-anomalies
participant Redis as Redis
participant SF as seed-forecasts (readInputKeys)
participant HC as /api/health
SI->>LTA: POST /api/infrastructure/v1/list-temporal-anomalies
Relay->>LTA: POST /api/infrastructure/v1/list-temporal-anomalies
LTA->>Redis: GET temporal:anomalies:v1
alt "cache hit (age < 900s)"
Redis-->>LTA: cached snapshot
LTA-->>SI: return cached (no seed-meta update)
else cache miss or stale
LTA->>LTA: compute fresh snapshot
LTA->>Redis: "SET temporal:anomalies:v1 (TTL=900s)"
Redis-->>LTA: "published=true"
LTA->>Redis: "SET seed-meta:temporal:anomalies {fetchedAt, recordCount} (TTL=604800s)"
LTA-->>SI: return snapshot
end
SF->>Redis: pipeline GET all forecast input keys
Redis-->>SF: parsedByKey
SF->>SF: buildForecastInputPresenceRows(parsedByKey)
SF->>SF: warnOnMissingForecastInputs(rows)
Note over SF: logs [ForecastInputs] warn per zero-record feed
HC->>Redis: GET temporal:anomalies:v1 strlen
HC->>Redis: GET seed-meta:temporal:anomalies
alt "data present AND seed-meta.fetchedAt < 45min ago"
HC-->>HC: "status=OK"
else missing or stale
HC-->>HC: "status=EMPTY or STALE_SEED (CRIT)"
end
Reviews (1): Last reviewed commit: "fix(forecast): surface missing forecast ..." | Re-trigger Greptile |
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
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.
Summary
Forecast runs now call out silent input-feed failures instead of letting missing or empty Redis inputs collapse into empty detector arrays with no operator signal. The run logs emit one
[ForecastInputs]warning per zero-record feed, including the Redis key andrecords=0, so ACLED, temporal anomalies, FRED series, and market-side inputs are visible when they no-fire.Forecast input coverage is also wired into health: temporal anomalies now require a strict data/meta signal, the forecast-specific FRED keys are registered with seed-meta expectations, and a successfully computed zero-anomaly temporal snapshot stays healthy rather than alarming as empty data.
Temporal anomalies are now kept warm through the relay warm-ping path and
seed-infra, and the producer writesseed-meta:temporal:anomalieswhen it refreshes the forecast input snapshot.Validation
./node_modules/.bin/tsx --test tests/forecast-input-observability.test.mjs tests/forecast-health-registration.test.mjs tests/relay-warm-ping-auth.test.mts tests/seed-warm-ping-origin.test.mjs tests/forecast-trace-export.test.mjs(339 pass)npm run typecheck:apigit diff --check origin/main...HEADnpm run typecheck,npm run typecheck:api, Convex typecheck, CJS syntax, Unicode safety, boundary/safe-html/Sentry/rate-limit/premium-fetch guards, edge bundle/tests, changed tests, version syncFixes #5101