detached d293fa6

feat: add TCP and HTTP readiness probes

TCP probe retries every 500ms until port accepts connections.
HTTP probe retries GET every 1s, expects 2xx/3xx.
Both are universal readiness checks for deploy pipeline.
commit
d293fa63b194deda7d288592a0b33b62a473935e
author
brice · · 2026-03-02 16:23 UTC
parent
867c255
signature
unsigned
2 files · +220 -0
Unified
UnifiedSplit
A internal/usecase/container/readiness.go +68 -0
@@ -0,0 +1,68 @@
1+package container
2+
3+import (
4+ "context"
5+ "fmt"
6+ "net"
7+ "net/http"
8+ "time"
9+)
10+
11+// tcpProbe attempts a TCP connection to addr, retrying every 500ms until
12+// success or timeout. This is the universal fallback readiness check —
13+// it verifies the process is at least accepting connections.
14+func tcpProbe(ctx context.Context, addr string, timeout time.Duration) error {
15+ deadline := time.Now().Add(timeout)
16+ for {
17+ if err := ctx.Err(); err != nil {
18+ return err
19+ }
20+ if time.Now().After(deadline) {
21+ return fmt.Errorf("TCP probe timeout after %s: %s not reachable", timeout, addr)
22+ }
23+
24+ conn, err := net.DialTimeout("tcp", addr, time.Second)
25+ if err == nil {
26+ conn.Close()
27+ return nil
28+ }
29+
30+ select {
31+ case <-time.After(500 * time.Millisecond):
32+ case <-ctx.Done():
33+ return ctx.Err()
34+ }
35+ }
36+}
37+
38+// httpProbe performs HTTP GET requests to url, retrying every 1s until a
39+// 2xx/3xx response or timeout. Used when gordon.health label is set.
40+func httpProbe(ctx context.Context, url string, timeout time.Duration) error {
41+ client := &http.Client{Timeout: 2 * time.Second}
42+ deadline := time.Now().Add(timeout)
43+ var lastStatus int
44+
45+ for {
46+ if err := ctx.Err(); err != nil {
47+ return err
48+ }
49+ if time.Now().After(deadline) {
50+ return fmt.Errorf("HTTP probe timeout after %s: last status %d from %s", timeout, lastStatus, url)
51+ }
52+
53+ resp, err := client.Get(url)
54+ if err == nil {
55+ lastStatus = resp.StatusCode
56+ resp.Body.Close()
57+ if lastStatus >= 200 && lastStatus < 400 {
58+ return nil
59+ }
60+ }
61+
62+ select {
63+ case <-time.After(time.Second):
64+ case <-ctx.Done():
65+ return ctx.Err()
66+ }
67+ }
68+}
@@ -0,0 +1,68 @@
1+package container
2+
3+import (
4+ "context"
5+ "fmt"
6+ "net"
7+ "net/http"
8+ "time"
9+)
10+
11+// tcpProbe attempts a TCP connection to addr, retrying every 500ms until
12+// success or timeout. This is the universal fallback readiness check —
13+// it verifies the process is at least accepting connections.
14+func tcpProbe(ctx context.Context, addr string, timeout time.Duration) error {
15+ deadline := time.Now().Add(timeout)
16+ for {
17+ if err := ctx.Err(); err != nil {
18+ return err
19+ }
20+ if time.Now().After(deadline) {
21+ return fmt.Errorf("TCP probe timeout after %s: %s not reachable", timeout, addr)
22+ }
23+
24+ conn, err := net.DialTimeout("tcp", addr, time.Second)
25+ if err == nil {
26+ conn.Close()
27+ return nil
28+ }
29+
30+ select {
31+ case <-time.After(500 * time.Millisecond):
32+ case <-ctx.Done():
33+ return ctx.Err()
34+ }
35+ }
36+}
37+
38+// httpProbe performs HTTP GET requests to url, retrying every 1s until a
39+// 2xx/3xx response or timeout. Used when gordon.health label is set.
40+func httpProbe(ctx context.Context, url string, timeout time.Duration) error {
41+ client := &http.Client{Timeout: 2 * time.Second}
42+ deadline := time.Now().Add(timeout)
43+ var lastStatus int
44+
45+ for {
46+ if err := ctx.Err(); err != nil {
47+ return err
48+ }
49+ if time.Now().After(deadline) {
50+ return fmt.Errorf("HTTP probe timeout after %s: last status %d from %s", timeout, lastStatus, url)
51+ }
52+
53+ resp, err := client.Get(url)
54+ if err == nil {
55+ lastStatus = resp.StatusCode
56+ resp.Body.Close()
57+ if lastStatus >= 200 && lastStatus < 400 {
58+ return nil
59+ }
60+ }
61+
62+ select {
63+ case <-time.After(time.Second):
64+ case <-ctx.Done():
65+ return ctx.Err()
66+ }
67+ }
68+}
A internal/usecase/container/readiness_test.go +152 -0
@@ -0,0 +1,152 @@
1+package container
2+
3+import (
4+ "context"
5+ "net"
6+ "net/http"
7+ "net/http/httptest"
8+ "sync/atomic"
9+ "testing"
10+ "time"
11+
12+ "github.com/stretchr/testify/assert"
13+ "github.com/stretchr/testify/require"
14+)
15+
16+// --- TCP Probe Tests ---
17+
18+func TestTCPProbe_Success(t *testing.T) {
19+ ln, err := net.Listen("tcp", "127.0.0.1:0")
20+ require.NoError(t, err)
21+ defer ln.Close()
22+
23+ // Accept connections in background so the probe succeeds
24+ go func() {
25+ for {
26+ conn, err := ln.Accept()
27+ if err != nil {
28+ return
29+ }
30+ conn.Close()
31+ }
32+ }()
33+
34+ ctx := testContext()
35+ err = tcpProbe(ctx, ln.Addr().String(), 5*time.Second)
36+ assert.NoError(t, err)
37+}
38+
39+func TestTCPProbe_Timeout(t *testing.T) {
40+ // Use a port that nothing listens on (port 1 is privileged, won't have a listener)
41+ ctx := testContext()
42+ err := tcpProbe(ctx, "127.0.0.1:1", 200*time.Millisecond)
43+ assert.Error(t, err)
44+ assert.Contains(t, err.Error(), "TCP probe timeout")
45+}
46+
47+func TestTCPProbe_ContextCancelled(t *testing.T) {
48+ ctx, cancel := context.WithCancel(testContext())
49+ cancel()
50+ err := tcpProbe(ctx, "127.0.0.1:1", 5*time.Second)
51+ assert.Error(t, err)
52+}
53+
54+func TestTCPProbe_DelayedListener(t *testing.T) {
55+ ln, err := net.Listen("tcp", "127.0.0.1:0")
56+ require.NoError(t, err)
57+ // Close initially so probe fails at first
58+ addr := ln.Addr().String()
59+ ln.Close()
60+
61+ // Re-open after 1 second
62+ go func() {
63+ time.Sleep(time.Second)
64+ newLn, err := net.Listen("tcp", addr)
65+ if err != nil {
66+ return
67+ }
68+ defer newLn.Close()
69+ for {
70+ conn, err := newLn.Accept()
71+ if err != nil {
72+ return
73+ }
74+ conn.Close()
75+ }
76+ }()
77+
78+ ctx := testContext()
79+ err = tcpProbe(ctx, addr, 5*time.Second)
80+ assert.NoError(t, err)
81+}
82+
83+// --- HTTP Probe Tests ---
84+
85+func TestHTTPProbe_Success(t *testing.T) {
86+ srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
87+ if r.URL.Path == "/healthz" {
88+ w.WriteHeader(200)
89+ return
90+ }
91+ w.WriteHeader(404)
92+ }))
93+ defer srv.Close()
94+
95+ ctx := testContext()
96+ err := httpProbe(ctx, srv.URL+"/healthz", 5*time.Second)
97+ assert.NoError(t, err)
98+}
99+
100+func TestHTTPProbe_ServerError_ThenSuccess(t *testing.T) {
101+ var calls int32
102+ srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
103+ n := atomic.AddInt32(&calls, 1)
104+ if n < 3 {
105+ w.WriteHeader(503)
106+ return
107+ }
108+ w.WriteHeader(200)
109+ }))
110+ defer srv.Close()
111+
112+ ctx := testContext()
113+ err := httpProbe(ctx, srv.URL+"/healthz", 10*time.Second)
114+ assert.NoError(t, err)
115+ assert.GreaterOrEqual(t, atomic.LoadInt32(&calls), int32(3))
116+}
117+
118+func TestHTTPProbe_Timeout(t *testing.T) {
119+ srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
120+ w.WriteHeader(503)
121+ }))
122+ defer srv.Close()
123+
124+ ctx := testContext()
125+ err := httpProbe(ctx, srv.URL+"/healthz", 500*time.Millisecond)
126+ assert.Error(t, err)
127+ assert.Contains(t, err.Error(), "HTTP probe timeout")
128+}
129+
130+func TestHTTPProbe_ContextCancelled(t *testing.T) {
131+ srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
132+ w.WriteHeader(503)
133+ }))
134+ defer srv.Close()
135+
136+ ctx, cancel := context.WithCancel(testContext())
137+ cancel()
138+ err := httpProbe(ctx, srv.URL+"/healthz", 5*time.Second)
139+ assert.Error(t, err)
140+}
141+
142+func TestHTTPProbe_RedirectIsSuccess(t *testing.T) {
143+ srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
144+ w.WriteHeader(302)
145+ }))
146+ defer srv.Close()
147+
148+ ctx := testContext()
149+ // Disable follow redirects for the test — the probe should treat 3xx as success
150+ err := httpProbe(ctx, srv.URL+"/healthz", 5*time.Second)
151+ assert.NoError(t, err)
152+}
@@ -0,0 +1,152 @@
1+package container
2+
3+import (
4+ "context"
5+ "net"
6+ "net/http"
7+ "net/http/httptest"
8+ "sync/atomic"
9+ "testing"
10+ "time"
11+
12+ "github.com/stretchr/testify/assert"
13+ "github.com/stretchr/testify/require"
14+)
15+
16+// --- TCP Probe Tests ---
17+
18+func TestTCPProbe_Success(t *testing.T) {
19+ ln, err := net.Listen("tcp", "127.0.0.1:0")
20+ require.NoError(t, err)
21+ defer ln.Close()
22+
23+ // Accept connections in background so the probe succeeds
24+ go func() {
25+ for {
26+ conn, err := ln.Accept()
27+ if err != nil {
28+ return
29+ }
30+ conn.Close()
31+ }
32+ }()
33+
34+ ctx := testContext()
35+ err = tcpProbe(ctx, ln.Addr().String(), 5*time.Second)
36+ assert.NoError(t, err)
37+}
38+
39+func TestTCPProbe_Timeout(t *testing.T) {
40+ // Use a port that nothing listens on (port 1 is privileged, won't have a listener)
41+ ctx := testContext()
42+ err := tcpProbe(ctx, "127.0.0.1:1", 200*time.Millisecond)
43+ assert.Error(t, err)
44+ assert.Contains(t, err.Error(), "TCP probe timeout")
45+}
46+
47+func TestTCPProbe_ContextCancelled(t *testing.T) {
48+ ctx, cancel := context.WithCancel(testContext())
49+ cancel()
50+ err := tcpProbe(ctx, "127.0.0.1:1", 5*time.Second)
51+ assert.Error(t, err)
52+}
53+
54+func TestTCPProbe_DelayedListener(t *testing.T) {
55+ ln, err := net.Listen("tcp", "127.0.0.1:0")
56+ require.NoError(t, err)
57+ // Close initially so probe fails at first
58+ addr := ln.Addr().String()
59+ ln.Close()
60+
61+ // Re-open after 1 second
62+ go func() {
63+ time.Sleep(time.Second)
64+ newLn, err := net.Listen("tcp", addr)
65+ if err != nil {
66+ return
67+ }
68+ defer newLn.Close()
69+ for {
70+ conn, err := newLn.Accept()
71+ if err != nil {
72+ return
73+ }
74+ conn.Close()
75+ }
76+ }()
77+
78+ ctx := testContext()
79+ err = tcpProbe(ctx, addr, 5*time.Second)
80+ assert.NoError(t, err)
81+}
82+
83+// --- HTTP Probe Tests ---
84+
85+func TestHTTPProbe_Success(t *testing.T) {
86+ srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
87+ if r.URL.Path == "/healthz" {
88+ w.WriteHeader(200)
89+ return
90+ }
91+ w.WriteHeader(404)
92+ }))
93+ defer srv.Close()
94+
95+ ctx := testContext()
96+ err := httpProbe(ctx, srv.URL+"/healthz", 5*time.Second)
97+ assert.NoError(t, err)
98+}
99+
100+func TestHTTPProbe_ServerError_ThenSuccess(t *testing.T) {
101+ var calls int32
102+ srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
103+ n := atomic.AddInt32(&calls, 1)
104+ if n < 3 {
105+ w.WriteHeader(503)
106+ return
107+ }
108+ w.WriteHeader(200)
109+ }))
110+ defer srv.Close()
111+
112+ ctx := testContext()
113+ err := httpProbe(ctx, srv.URL+"/healthz", 10*time.Second)
114+ assert.NoError(t, err)
115+ assert.GreaterOrEqual(t, atomic.LoadInt32(&calls), int32(3))
116+}
117+
118+func TestHTTPProbe_Timeout(t *testing.T) {
119+ srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
120+ w.WriteHeader(503)
121+ }))
122+ defer srv.Close()
123+
124+ ctx := testContext()
125+ err := httpProbe(ctx, srv.URL+"/healthz", 500*time.Millisecond)
126+ assert.Error(t, err)
127+ assert.Contains(t, err.Error(), "HTTP probe timeout")
128+}
129+
130+func TestHTTPProbe_ContextCancelled(t *testing.T) {
131+ srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
132+ w.WriteHeader(503)
133+ }))
134+ defer srv.Close()
135+
136+ ctx, cancel := context.WithCancel(testContext())
137+ cancel()
138+ err := httpProbe(ctx, srv.URL+"/healthz", 5*time.Second)
139+ assert.Error(t, err)
140+}
141+
142+func TestHTTPProbe_RedirectIsSuccess(t *testing.T) {
143+ srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
144+ w.WriteHeader(302)
145+ }))
146+ defer srv.Close()
147+
148+ ctx := testContext()
149+ // Disable follow redirects for the test — the probe should treat 3xx as success
150+ err := httpProbe(ctx, srv.URL+"/healthz", 5*time.Second)
151+ assert.NoError(t, err)
152+}

Keyboard shortcuts

Single-key shortcuts are on. ctrl k and esc work whatever this says.

ctrl k :
go to a section, branch, tag, file or repository
?
this list
j k
move the row cursor down / up
enter
open the row under the cursor; on Files, a directory opens or closes in place
↑ ↓ ← → home end
in the file tree (focus on a row): previous / next row, close / open a directory or go to its parent / first entry, first / last row. They never open a file
l h
on Files: open a directory or a file / close a directory or go up to its parent, then to the parent page. Elsewhere: into the side panel (on a merge request, its files) / back to the content; there j k move between its controls
g g G
first / last row
/
filter the list; on Files, find a file anywhere in the repository (esc clears, enter opens the best match)
b
switch branch or tag
g s g f g c g b g m g i g r g a
tabs: summary, files, commits, branches & tags, merge requests, issues, releases, actions (when available)
g h
home
w
wide mode: use the whole width of the screen, or not (remembered)
y
copy a permalink pinned to the commit
d u x
diff: split / unified / collapse or expand all
e E
diff: the whole file in view / every file, with the lines the hunks hide
n
merge request: next note in the diff
esc
close a dialog, leave a field

Branches 7

Tags 106

↑↓ moveenter openesc close