Alerting Concepts
Reference for every field in the alert configuration form. Building a rule now? Use Create and manage alerts.
Skim the condition sentence below, then jump to the field you are filling in the
form. Headings here are stable deep links from portal (i) help.
Condition as a sentence
An alert is one sentence:
If this metric (for this app, with optional filters) over this time window, compared with this threshold, then notify these channels.
| # | Field | Jump |
|---|---|---|
| 1 | Application | Application |
| 2 | Category & metric | Category and metric |
| 3 | Aggregator | Aggregator |
| 4 | Operator | Operator |
| 5 | Threshold | Threshold |
| 6 | Time window | Time window |
| 7 | Filters (optional) | Filters |
| 8 | Severity | Severity |
| 9 | Channels | Notification channels |
| 10 | Enabled / Disabled | Status |
Application
Alerts watch traffic for a specific application. Pick the app whose API or webhook activity you want to monitor.
- Open the alert configuration form.
- Choose the application from the list.
- All metrics and filters apply only to that app’s traffic.
Category and metric
Categories available for monitoring today
| Category | Use when |
|---|---|
| API | You care about requests to your APIs |
| Webhook | You care about outbound webhook deliveries |
Metrics by category
- API metrics
- Webhook metrics
| Metric | What it tells you |
|---|---|
| API request count | How many requests arrived |
| API error count | How many requests failed |
| API error rate | What share of requests failed |
| API success rate | What share of requests succeeded |
| API delivery latency | How long responses took |
| API payload size (average) | How large request payloads were |
| Rate-limit exceeded count | How often responses looked rate-limited |
| Metric | What it tells you |
|---|---|
| Webhook delivery count | How many deliveries were attempted |
| Webhook delivery success rate | What share succeeded |
| Webhook delivery failure rate | What share failed |
| Webhook retry count | How many deliveries needed retries |
| Webhook pending count | How many are still pending |
| Webhook payload size (average) | How large webhook payloads were |
Aggregator
An alert condition answers three questions:
- How to combine recent values (aggregator)
- How to compare (operator)
- What limit matters (threshold)
ChainIT looks at minute-by-minute measurements inside your time window, then combines them:
| Option | Meaning | Good for |
|---|---|---|
| Maximum | Highest value in the window | Catching a spike in any single minute |
| Minimum | Lowest value in the window | Catching a drop |
| Average | Typical value across the window | Smooth rate or latency trends |
| Count | How many minute samples exist | Sample coverage, not total request volume |
The portal only shows combinations that are valid for the metric you selected.
Aggregator Count counts minute samples, not how many API calls happened. For traffic volume, prefer Maximum or Average.
Operator
| Operator | Meaning |
|---|---|
Greater than (>) | Fire when the result is above the threshold |
Less than (<) | Fire when the result is below the threshold |
Greater than or equal (>=) | Fire at or above the threshold |
Less than or equal (<=) | Fire at or below the threshold |
Equal (==) | Fire when exactly equal |
Not equal (!=) | Fire when not equal |
Threshold
The threshold is the number you care about.
| Goal | Typical setup |
|---|---|
| Peak minute over 1000 requests | Metric = request count, combine with Maximum, > 1000 |
| Average over 200 requests/minute | Metric = request count, combine with Average, > 200 |
| Error rate above 10% | Metric = error rate, combine with Average, > 0.10 |
| Latency above 500 ms | Metric = latency, combine with Average, > 500 |
For rate metrics, a threshold of 15% is entered as 0.15, not
15.
Time window
How far back the rule looks:
| Option | Meaning |
|---|---|
| 15 minutes | Short, responsive monitoring |
| 30 minutes | Medium window |
| 45 minutes | Longer trend window |
The same window also sets how long notification cooldown lasts after a fire.
Filters
Filters narrow which traffic is included. Use them when the whole app is too broad.
Common filter types:
- HTTP methods (GET, POST, and similar)
- Endpoints / paths
- Status codes
- Webhook subscriptions or events
- Payload size or latency bounds
Severity
Severity is a triage label. It does not change the math of the rule.
| Severity | Typical use |
|---|---|
| Info / Debug | Awareness |
| Warning | Needs attention soon |
| Error | Customer impact likely |
| Critical / Emergency | Act immediately |
Notification channels
Notification channels are where alert messages go: email and/or webhook.
- Create channels under Alerting / Notification channels.
- Attach one or more channels when you save an alert configuration.
- Reuse the same channel across many rules.
Portal steps: Create and manage: Step 1.
Status: Enabled vs Disabled
| Status | Meaning |
|---|---|
| Enabled | The rule is actively monitored |
| Disabled | The rule is paused and will not fire |
Disable a rule when you want to pause monitoring without deleting it.
How the pieces connect
Next steps
- Create and manage alerts
build a channel and rule
- Overview
end-to-end flow and monitoring summary