Monitoring
Monitoring answers one question: is the data still flowing, and is it still good? This page covers what Signals shows you today and, just as importantly, what it does not.
What is recorded
Section titled “What is recorded”Delivery counts are recorded for every destination and for batch sources, not just for Meta.
| Recorded | What it tells you |
|---|---|
| Events received | What arrived from your source. |
| Events delivered | What was successfully transmitted to the destination. |
| Events failed or skipped | What did not make it, with the error message that explains why. |
The comparison between received and delivered is the useful one. Delivered consistently below received means events are being lost between receipt and delivery.
Gateway reporting
Section titled “Gateway reporting”The richer per-identifier reporting is scoped to web conversions through the Meta Conversions API Gateway. Once Gateway setup is complete, the Signal count icon opens it.
| Metric | What it shows |
|---|---|
| Total Incoming Signals | The total number of events received by the gateway. |
| Total Outgoing Signals | The number of events successfully transmitted to Meta. |
| EMQ Score | Event Match Quality for the events sent, refreshed on a daily pull. |
| Event Volume Overview | Event volume over time, with a range of last 7 days, last 30 days or custom, and incoming and outgoing totals for the period. |
| Event-Level Breakdown | A bar chart by event type, with filtering and highlighting of specific events. |
| EMQ Table | EMQ per event type, parameter coverage percentage, and a trend indicator of improving, stable or declining. |
Datahash’s guidance is to compare Total Incoming Signals with Total Outgoing Signals to assess transmission efficiency, and to pay attention to events with low EMQ scores, since those limit advertising effectiveness.
Event freshness
Section titled “Event freshness”Meta reports freshness separately, in Events Manager under Event Details, on the Event Freshness tab. It compares event_time with the time Meta received the event. Delay between the two causes attribution errors, so it is worth checking alongside volume.
There is no alerting
Section titled “There is no alerting”Signals does not alert you when something breaks. There are no channels, no recipients, no thresholds and no escalation, and nothing fires automatically.
Nor is there a customer facing delivery log, an error catalog, or a per-event receipt showing what each destination accepted or rejected. Failure counts and error messages are recorded in reporting, but you have to go and look at them.
What to check yourself
Section titled “What to check yourself”- Compare received and delivered counts for each connection.
- Check event volume in each platform’s own event manager against what you expect.
- Check match quality per event type, and investigate any event that drops.
- Check deduplication overlap where you run both a pixel and a server connection.
- Watch for a sudden change after any source schema change, since a renamed column stops being recognized without an error.
A weekly glance at the first two catches most problems while they are still small.