Pim van Pelt
1815675fb6
Add StateDisabled for operator-initiated disable, keeping StateRemoved for backends that disappear during a config reload. Previously both used StateRemoved, which was confusing: "removed" implies the backend no longer exists in config, but a disabled backend is still present and can be re-enabled on the fly. - health: add StateDisabled with String() "disabled", Disable() method with probe code "disabled". Record() rejects probes in all three inactive states (paused, disabled, removed). - checker: DisableBackend calls backend.Disable() instead of Remove(). - docs: healthchecks.md rewritten for pause (goroutine cancelled, not just results discarded), and separate disabled/removed state rows. user-guide.md updated to match. - Makefile: add fixstyle target (gofmt -w .).
169 lines
6.6 KiB
Plaintext
169 lines
6.6 KiB
Plaintext
*** Settings ***
|
|
Library OperatingSystem
|
|
Resource ../common.robot
|
|
|
|
Suite Setup Setup Suite
|
|
Suite Teardown Cleanup Suite
|
|
|
|
|
|
*** Variables ***
|
|
${lab-name} maglevd-test
|
|
${lab-file} maglevd-lab/maglevd.clab.yml
|
|
${runtime} docker
|
|
${MAGLEVD_NODE} clab-maglevd-test-maglevd
|
|
${METRICS_URL} http://172.20.30.2:9091/metrics
|
|
|
|
|
|
*** Test Cases ***
|
|
Deploy maglevd-test lab
|
|
[Documentation] Deploy the containerlab topology. The maglevd node starts
|
|
... automatically as PID 1 via start.sh and begins probing the nginx
|
|
... backends immediately.
|
|
${rc} ${output} = Run And Return Rc And Output
|
|
... ${CLAB_BIN} --runtime ${runtime} deploy -t ${CURDIR}/${lab-file}
|
|
Log ${output}
|
|
Should Be Equal As Integers ${rc} 0
|
|
Sleep 3s Wait for nginx containers and probes to converge
|
|
|
|
All backends reach up state
|
|
[Template] Backend Should Be Up
|
|
nginx1
|
|
nginx2
|
|
nginx3
|
|
|
|
Health checks are reaching all backends
|
|
[Template] Probe Count Should Be Positive
|
|
nginx1
|
|
nginx2
|
|
nginx3
|
|
|
|
Pause backend stops probing
|
|
Maglevc set backend nginx1 pause
|
|
Backend Should Have State nginx1 paused
|
|
Sleep 1s
|
|
${before} = Get Probe Count nginx1
|
|
Sleep 2s Wait to confirm no new probes arrive
|
|
${after} = Get Probe Count nginx1
|
|
Should Be True ${after} == ${before}
|
|
... Probe count for nginx1 grew while paused: ${before} → ${after}
|
|
|
|
Resume backend restarts probing
|
|
Maglevc set backend nginx1 resume
|
|
${before} = Get Probe Count nginx1
|
|
Sleep 2s Wait for resumed probes to accumulate
|
|
${after} = Get Probe Count nginx1
|
|
Should Be True ${after} > ${before}
|
|
... Probe count for nginx1 did not grow after resume: ${before} → ${after}
|
|
Wait Until Keyword Succeeds 5s 500ms
|
|
... Backend Should Be Up nginx1
|
|
|
|
Disable backend stops probing
|
|
Maglevc set backend nginx2 disable
|
|
Backend Should Have State nginx2 disabled
|
|
Backend Should Be Disabled nginx2
|
|
Sleep 1s
|
|
${before} = Get Probe Count nginx2
|
|
Sleep 2s Wait to confirm probes stopped
|
|
${after} = Get Probe Count nginx2
|
|
Should Be True ${after} == ${before}
|
|
... Probe count for nginx2 grew while disabled: ${before} → ${after}
|
|
|
|
Enable backend restarts probing
|
|
Maglevc set backend nginx2 enable
|
|
${before} = Get Probe Count nginx2
|
|
Sleep 2s Wait for re-enabled probes to accumulate
|
|
${after} = Get Probe Count nginx2
|
|
Should Be True ${after} > ${before}
|
|
... Probe count for nginx2 did not grow after enable: ${before} → ${after}
|
|
Wait Until Keyword Succeeds 5s 500ms
|
|
... Backend Should Be Up nginx2
|
|
|
|
|
|
Prometheus endpoint is reachable
|
|
${rc} ${output} = Run And Return Rc And Output
|
|
... curl -sf ${METRICS_URL}
|
|
Log ${output}
|
|
Should Be Equal As Integers ${rc} 0
|
|
Should Contain ${output} maglev_backend_state
|
|
|
|
Prometheus reports all backends up
|
|
${output} = Scrape Metrics
|
|
# Each backend should have state="up" = 1.
|
|
Should Contain ${output} maglev_backend_state{address="172.20.30.11",backend="nginx1",healthcheck="http-check",state="up"} 1
|
|
Should Contain ${output} maglev_backend_state{address="172.20.30.12",backend="nginx2",healthcheck="http-check",state="up"} 1
|
|
Should Contain ${output} maglev_backend_state{address="172.20.30.13",backend="nginx3",healthcheck="http-check",state="up"} 1
|
|
|
|
Prometheus reports probe counters
|
|
${output} = Scrape Metrics
|
|
Should Match Regexp ${output} maglev_probe_total\\{backend="nginx1".*result="success".*\\}\\s+[1-9]
|
|
Should Match Regexp ${output} maglev_probe_total\\{backend="nginx2".*result="success".*\\}\\s+[1-9]
|
|
Should Match Regexp ${output} maglev_probe_total\\{backend="nginx3".*result="success".*\\}\\s+[1-9]
|
|
|
|
Prometheus reports probe duration histogram
|
|
${output} = Scrape Metrics
|
|
Should Match Regexp ${output} maglev_probe_duration_seconds_count\\{backend="nginx1".*\\}\\s+[1-9]
|
|
|
|
Prometheus reports pool weights
|
|
${output} = Scrape Metrics
|
|
Should Contain ${output} maglev_frontend_pool_backend_weight{backend="nginx1",frontend="http-vip",pool="primary"} 100
|
|
Should Contain ${output} maglev_frontend_pool_backend_weight{backend="nginx3",frontend="http-vip",pool="fallback"} 100
|
|
|
|
Prometheus reports transition counters
|
|
${output} = Scrape Metrics
|
|
# All backends transitioned unknown → up during startup.
|
|
Should Match Regexp ${output} maglev_backend_transitions_total\\{backend="nginx1",from="unknown",to="up"\\}\\s+[1-9]
|
|
|
|
|
|
*** Keywords ***
|
|
Setup Suite
|
|
${arch} = Run go env GOARCH
|
|
Set Suite Variable ${ARCH} ${arch}
|
|
|
|
Cleanup Suite
|
|
Run docker logs ${MAGLEVD_NODE} > ${EXECDIR}/tests/out/maglevd.log 2>&1
|
|
Run ${CLAB_BIN} --runtime ${runtime} destroy -t ${CURDIR}/${lab-file} --cleanup
|
|
|
|
Maglevc
|
|
[Documentation] Run a maglevc command inside the maglevd container.
|
|
[Arguments] ${cmd}
|
|
${rc} ${output} = Run And Return Rc And Output
|
|
... docker exec ${MAGLEVD_NODE} /opt/maglev/build/${ARCH}/maglevc --color\=false ${cmd}
|
|
Log ${output}
|
|
Should Be Equal As Integers ${rc} 0
|
|
RETURN ${output}
|
|
|
|
Backend Should Be Up
|
|
[Arguments] ${name}
|
|
${output} = Maglevc show backends ${name}
|
|
Should Match Regexp ${output} state\\s+up
|
|
|
|
Backend Should Have State
|
|
[Arguments] ${name} ${expected_state}
|
|
${output} = Maglevc show backends ${name}
|
|
Should Match Regexp ${output} state\\s+${expected_state}
|
|
|
|
Backend Should Be Disabled
|
|
[Arguments] ${name}
|
|
${output} = Maglevc show backends ${name}
|
|
Should Match Regexp ${output} enabled\\s+false
|
|
|
|
Get Probe Count
|
|
[Documentation] Return the number of HTTP health-check requests seen in a backend's nginx log.
|
|
[Arguments] ${name}
|
|
${output} = Run docker logs clab-${lab-name}-${name} 2>/dev/null | grep -c "GET /" || echo 0
|
|
${count} = Convert To Integer ${output.strip()}
|
|
RETURN ${count}
|
|
|
|
Probe Count Should Be Positive
|
|
[Arguments] ${name}
|
|
${count} = Get Probe Count ${name}
|
|
Should Be True ${count} > 0
|
|
... No health-check requests found in nginx logs for ${name}
|
|
|
|
Scrape Metrics
|
|
[Documentation] Fetch the Prometheus /metrics endpoint from the maglevd container.
|
|
${rc} ${output} = Run And Return Rc And Output
|
|
... curl -sf ${METRICS_URL}
|
|
Should Be Equal As Integers ${rc} 0
|
|
RETURN ${output}
|