Monitors

A monitor defines what to check and how often. UpSlim supports two monitor types: http and tcp.

Common fields

These fields apply to both HTTP and TCP monitors.

Field	Type	Default	Description
name	string	required	Unique name for this monitor
type	http | tcp	required	Monitor type
interval	duration	60s	How often to run the check
timeout	duration	30s	Maximum time to wait
failure_threshold	integer	3	Consecutive failures before alerting
success_threshold	integer	2	Consecutive successes to mark recovered
send_on_resolved	boolean	true	Send recovery notification
conditions	list	required	Conditions that must all pass
alerts	list	[]	Alert providers to notify

HTTP monitor

Performs an HTTP request and evaluates the response.

monitors:
  - name: api-health
    type: http
    url: "https://api.example.com/health"
    interval: 30s
    timeout: 10s
    method: GET                           # optional, default: GET
    headers:
      Authorization: "Bearer ${API_TOKEN}"
    conditions:
      - "[STATUS] == 200"
      - "[RESPONSE_TIME] < 500"
      - "[BODY].status == healthy"
    alerts:
      - name: slack-ops

HTTP-specific fields

Field	Type	Default	Description
url	string	required	Full URL including scheme
method	string	GET	HTTP method
headers	map	{}	Request headers
body	string	none	Request body (for POST/PUT)

Available conditions for HTTP

Expression	Description
[STATUS] == 200	HTTP status code equals 200
[STATUS] < 400	HTTP status code less than 400
[RESPONSE_TIME] < 500	Response time in milliseconds
[BODY] == ok	Raw response body equals string
[BODY].field == value	JSON body dot-path equals value

See the Conditions DSL for the full syntax.

TCP monitor

Opens a TCP connection and checks it succeeds within the timeout.

monitors:
  - name: postgres
    type: tcp
    host: "db.internal"
    port: 5432
    interval: 60s
    timeout: 5s
    conditions:
      - "[CONNECTED] == true"
    alerts:
      - name: slack-ops

TCP-specific fields

Field	Type	Default	Description
host	string	required	Hostname or IP address
port	integer	required	Port number (1–65535)

Available conditions for TCP

Expression	Description
[CONNECTED] == true	TCP handshake completed successfully
[RESPONSE_TIME] < 100	Time to establish connection in ms

Per-monitor alert overrides

Thresholds can be overridden per alert reference, not just at the monitor level:

monitors:
  - name: critical-api
    type: http
    url: "https://payments.example.com/health"
    conditions:
      - "[STATUS] == 200"
    failure_threshold: 3    # monitor-level default
    alerts:
      - name: slack-ops
        failure_threshold: 1  # alert fires after 1 failure for this provider

WARNING

Per-alert failure_threshold overrides are read by the alert state machine. The monitor still runs every interval, but the alert fires only when the overridden threshold is reached.

Multiple monitors example

defaults:
  interval: 60s
  timeout: 10s

monitors:
  - name: web
    type: http
    url: "https://example.com"
    conditions:
      - "[STATUS] == 200"
      - "[RESPONSE_TIME] < 2000"

  - name: api
    type: http
    url: "https://api.example.com/health"
    interval: 30s        # override global default
    conditions:
      - "[STATUS] == 200"
      - "[BODY].status == healthy"

  - name: database
    type: tcp
    host: "db.internal"
    port: 5432
    conditions:
      - "[CONNECTED] == true"

  - name: cache
    type: tcp
    host: "redis.internal"
    port: 6379
    conditions:
      - "[CONNECTED] == true"