ptr.pet
at protocol indexer with flexible filtering, xrpc queries, and a cursor-backed event stream, built on fjall
at-protocol
atproto
indexer
rust
fjall
[tests] add throttling integration test
+1
AGENTS.md
+1
AGENTS.md
···
88
88
### Testing
89
89
- `nu tests/repo_sync_integrity.nu` - Runs the full integration test suite using Nushell. This builds the binary, starts a temporary instance, performs a backfill against a real PDS, and verifies record integrity.
90
90
- `nu tests/verify_crawler.nu` - Verifies full-network crawler functionality using a mock relay.
91
+
- `nu tests/throttling_test.nu` - Verifies crawler throttling logic when pending queue is full.
91
92
- `nu tests/stream_test.nu` - Tests WebSocket streaming functionality. Verifies both live event streaming during backfill and historical replay with cursor.
92
93
- `nu tests/authenticated_stream_test.nu` - Tests authenticated event streaming. Verifies that create, update, and delete actions on a real account are correctly streamed by Hydrant in the correct order. Requires `TEST_REPO` and `TEST_PASSWORD` in `.env`.
93
94
- `nu tests/debug_endpoints.nu` - Tests debug/introspection endpoints (`/debug/iter`, `/debug/get`) and verifies DB content and serialization.
+148
tests/throttling_test.nu
+148
tests/throttling_test.nu
···
1
+
#!/usr/bin/env nu
2
+
use common.nu *
3
+
4
+
def main [] {
5
+
# 1. ensure http-nu is installed
6
+
if (which http-nu | is-empty) {
7
+
print "http-nu not found, installing..."
8
+
cargo install http-nu
9
+
}
10
+
11
+
# 2. setup ports and paths
12
+
let port = 3010
13
+
let mock_port = 3012
14
+
let url = $"http://localhost:($port)"
15
+
let mock_url = $"http://localhost:($mock_port)"
16
+
let db_path = (mktemp -d -t hydrant_throttling.XXXXXX)
17
+
18
+
print $"testing crawler throttling..."
19
+
print $"database path: ($db_path)"
20
+
21
+
# 3. start mock relay
22
+
print $"starting mock relay on ($mock_port)..."
23
+
let mock_pid = (
24
+
bash -c $"http-nu :($mock_port) tests/mock_relay.nu > ($db_path)/mock.log 2>&1 & echo $!"
25
+
| str trim
26
+
| into int
27
+
)
28
+
print $"mock relay pid: ($mock_pid)"
29
+
30
+
# give mock relay a moment
31
+
sleep 1sec
32
+
33
+
# 4. start hydrant with low throttling limits
34
+
let binary = build-hydrant
35
+
36
+
let log_file = $"($db_path)/hydrant.log"
37
+
print $"starting hydrant - logs at ($log_file)..."
38
+
39
+
let hydrant_pid = (
40
+
with-env {
41
+
HYDRANT_DATABASE_PATH: ($db_path),
42
+
HYDRANT_FULL_NETWORK: "true",
43
+
HYDRANT_RELAY_HOST: ($mock_url),
44
+
HYDRANT_DISABLE_FIREHOSE: "true",
45
+
HYDRANT_DISABLE_BACKFILL: "true", # disable backfill so pending count stays up
46
+
HYDRANT_API_PORT: ($port | into string),
47
+
HYDRANT_LOG_LEVEL: "debug",
48
+
HYDRANT_CRAWLER_MAX_PENDING_REPOS: "2",
49
+
HYDRANT_CRAWLER_RESUME_PENDING_REPOS: "1"
50
+
} {
51
+
sh -c $"($binary) >($log_file) 2>&1 & echo $!" | str trim | into int
52
+
}
53
+
)
54
+
print $"hydrant started with pid: ($hydrant_pid)"
55
+
56
+
mut success = false
57
+
58
+
try {
59
+
if (wait-for-api $url) {
60
+
print "hydrant api is up."
61
+
62
+
# wait for crawler to run and hit limit
63
+
print "waiting for crawler to hit throttling limit..."
64
+
65
+
# retry check for 30s
66
+
for i in 1..30 {
67
+
let stats = (http get $"($url)/stats?accurate=true").counts
68
+
let pending = ($stats.pending | into int)
69
+
70
+
# we expect 5 repos from the mock, but max pending is 2.
71
+
# wait, the crawler fetches a page (5 repos) THEN adds to DB.
72
+
# so pending will jump to 5.
73
+
# then next loop, it checks pending > 2.
74
+
# so pending should be 5.
75
+
76
+
print $"[($i)/30] pending: ($pending)"
77
+
78
+
if $pending >= 5 {
79
+
print "crawler discovered repos."
80
+
break
81
+
}
82
+
83
+
sleep 1sec
84
+
}
85
+
86
+
# now check logs for throttling message
87
+
print "checking logs for throttling message..."
88
+
sleep 2sec # give logging a moment
89
+
90
+
let logs = (open $log_file | str replace --all "\n" " ")
91
+
if ($logs | str contains "crawler throttling: pending repos") {
92
+
print "CONFIRMED: crawler is throttling!"
93
+
94
+
# now testing resumption
95
+
print "testing resumption by removing repos..."
96
+
97
+
# remove 4 repos to drop pending (5) to 1 (<= resume limit 1)
98
+
# mock repos are did:web:mock1.com ... mock5.com
99
+
let to_remove = {
100
+
dids: [
101
+
"did:web:mock1.com",
102
+
"did:web:mock2.com",
103
+
"did:web:mock3.com",
104
+
"did:web:mock4.com"
105
+
]
106
+
}
107
+
108
+
http post --content-type application/json $"($url)/repo/remove" $to_remove
109
+
110
+
print "waiting for crawler to wake up (max 10s)..."
111
+
sleep 15sec
112
+
113
+
# check logs for resumption message
114
+
let logs_after = (open $log_file | str replace --all "\n" " ")
115
+
if ($logs_after | str contains "crawler resuming") {
116
+
print "CONFIRMED: crawler resumed!"
117
+
$success = true
118
+
} else {
119
+
print "FAILED: resumption message not found in logs"
120
+
$success = false
121
+
}
122
+
123
+
} else {
124
+
print "FAILED: throttling message not found in logs"
125
+
}
126
+
127
+
} else {
128
+
print "hydrant failed to start."
129
+
}
130
+
} catch { |e|
131
+
print $"test failed with error: ($e)"
132
+
}
133
+
134
+
# cleanup
135
+
print "stopping processes..."
136
+
try { kill $hydrant_pid }
137
+
try { kill $mock_pid }
138
+
139
+
if $success {
140
+
print "test passed!"
141
+
exit 0
142
+
} else {
143
+
print "test failed!"
144
+
print "hydrant logs:"
145
+
open $log_file | tail -n 20
146
+
exit 1
147
+
}
148
+
}