llm-proxy/configs/fuzz.yml at main · Instawork/llm-proxy · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
# Fuzz / fake-upstream overlay (ENVIRONMENT=fuzz).
# Requires LLM_PROXY_ALLOW_FAKE_MODE=1 and LLM_PROXY_ALLOW_TEST_MODE=1 in process env.

enabled: true

features:
  fake_upstream:
    enabled: true
    chaos_failure_rate: 0.15
    chaos_seed: 0
    latency_ms: 0
    jitter_ms: 0
  cost_tracking:
    enabled: true
    async: true
    workers: 3
    queue_size: 50
    flush_interval: 2
    transports:
      - type: "file"
        file:
          path: "./logs/cost-tracking.jsonl"
  api_key_management:
    enabled: true
    table_name: "llm-proxy-api-keys-fuzz"
    region: "us-west-2"
    key_prefix: "iw"
    auto_create_table: true
    provisioning:
      enabled: false
  circuit_breaker:
    enabled: true
    mode: enforce
    backend: "redis"
    redis:
      address: "redis:6379"
      db: 5
    failure_threshold: 3
    window_seconds: 60
    cooldown_seconds: 10
    max_transient_retries: 2
    max_rate_limit_retries: 2
    global_rate_limit_escalation_window: 60
    retry_contribution_mode: "log"
    test_mode_enabled: true
  admin_dashboard:
    enabled: true
    dev_bypass_login: true
    users:
      dynamodb:
        table_name: "llm-proxy-admin-users-fuzz"
        region: "us-west-2"
        auto_create_table: true
    rollups:
      enabled: true
      redis:
        address: "redis:6379"
        db: 6
      retention_days: 7
      history_days: 7
  pii_redact:
    enabled: true
    analyzer_url: "http://presidio:3000"
    fail_mode: "open"
    allow_per_key_override: true
    wire_placeholders: true
  redact_api:
    enabled: false
  rate_limiting:
    enabled: true
    backend: "redis"
    estimation:
      max_sample_bytes: 200000
      bytes_per_token: 4
      chars_per_token: 4
      provider_chars_per_token:
        openai: 5
        anthropic: 3
    limits:
      requests_per_minute: 0
      tokens_per_minute: 0
      requests_per_day: 0
      tokens_per_day: 0
    redis:
      address: "redis:6379"
      password: ""
      db: 4
  history:
    backend: local
    role: global
    streams: [cost, usage, ratelimit]
    max_records: 100
    max_bytes: 1048576
    max_age_seconds: 60
    gzip: true
    local:
      dir: logs/history

providers:
  bedrock:
    enabled: false