Generated reference · Performance

Performance reference

Latency budgets and load thresholds — the CI-enforced targets a change must not regress, parsed from the benchmark + k6 suites.

14budgets17threshold sets147checks
testing/performance
On this page

The platform’s performance budgets — the latency targets and load thresholds a change must not regress, generated from the committed benchmark + k6 suites under testing/performance/. Each is asserted by CI, so this is the enforced contract, not aspiration.

14Latency budgets
17k6 threshold sets
7Load-tested domains
147Threshold checks
Scope: declared, CI-enforced budgets/thresholds — the targets a change must hold — not live or representative measured results (those vary per run and are not committed). Product game-engine frame budgets live in their product space.

Latency budgets (14)#

Per-hop p50/p95/p99 targets (milliseconds) for cross-domain calls and health checks, asserted by the benchmark suites.

yemaya-health#
/healthp50 50ms p95 100ms p99 200ms
lilith-health#
/healthp50 50ms p95 100ms p99 200ms
isis-health#
/healthp50 50ms p95 100ms p99 200ms
sophia-health#
/healthp50 50ms p95 100ms p99 200ms
hathor-health#
/healthp50 50ms p95 100ms p99 200ms
bellona-health#
/healthp50 50ms p95 100ms p99 200ms
lilith-bff-to-isis-enqueuelilith → isis#
POST /api/v1/generationsp50 150ms p95 400ms p99 800ms
lilith-bff-to-sophia-asklilith → sophia#
POST /api/v1/askp50 300ms p95 800ms p99 1500ms
hathor-to-sophia-resolve-entityhathor → sophia#
POST /api/v1/entities/resolvep50 120ms p95 300ms p99 600ms
hathor-to-bellona-world-exporthathor → bellona#
POST /v1/buildsp50 200ms p95 500ms p99 1000ms
isis-to-bellona-asset-exportisis → bellona#
POST /v1/exportsp50 150ms p95 400ms p99 900ms
yemaya-to-bellona-session-syncyemaya → bellona#
POST /v1/sessionsp50 100ms p95 250ms p99 500ms

Load thresholds (k6) (17)#

aphrodite

combined-load-test#
http_req_durationp(95)<1500
http_req_failedrate<0.05
errorsrate<0.05
combined_latencyp(95)<1000
streaming_latencyp(95)<500
payment_latencyp(95)<2000
device_latencyp(95)<500
chat_latencyp(95)<300
total_operationscount>10000
device-flooding#
http_req_durationp(95)<1000
http_req_failedrate<0.10
errorsrate<0.10
vibrate_command_latencyp(95)<300
stop_command_latencyp(95)<200
pattern_command_latencyp(95)<400
batch_command_latencyp(95)<800
tip_trigger_latencyp(95)<500
commands_succeededcount>5000
tip-storm#
http_req_durationp(95)<2000
http_req_failedrate<0.05
errorsrate<0.05
tip_processing_latencyp(95)<1500
device_trigger_latencyp(95)<500
balance_check_latencyp(95)<200
successful_tipscount>1000
viewer-simulation#
http_req_durationp(95)<1000
http_req_failedrate<0.05
errorsrate<0.05
stream_discovery_latencyp(95)<500
stream_join_latencyp(95)<1000
chat_message_latencyp(95)<300
streaming_latencyp(95)<500
payment_latencyp(95)<2000

bellona

BRIDGE_THRESHOLDS#
http_req_failedrate<0.02
session_start_success_raterate>0.98
build_start_success_raterate>0.98
export_start_success_raterate>0.98
build_status_success_raterate>0.99
engine_health_success_raterate>0.99
bellona_session_start_latencyp(95)<250p(99)<500
bellona_build_start_latencyp(95)<500p(99)<1000
bellona_build_status_latencyp(95)<150p(99)<300
bellona_build_logs_latencyp(95)<500
bellona_export_start_latencyp(95)<400p(99)<800
bellona_export_status_latencyp(95)<200
bellona_engine_health_latencyp(95)<100p(99)<250
errorsrate<0.05

hathor

NARRATIVE_THRESHOLDS#
http_req_failedrate<0.02
quest_success_raterate>0.95
hathor_quest_create_latencyp(95)<3500p(99)<7000
errorsrate<0.05
SIMULATION_THRESHOLDS#
http_req_failedrate<0.02
simulation_enqueue_success_raterate>0.98
simulation_first_frame_success_raterate>0.95
hathor_simulation_enqueue_latencyp(95)<500p(99)<1000
hathor_simulation_first_frame_latencyp(95)<1500p(99)<3000
errorsrate<0.05
WORLD_API_THRESHOLDS#
http_req_failedrate<0.02
create_world_success_raterate>0.98
export_world_success_raterate>0.95
hathor_create_world_latencyp(95)<400p(99)<800
hathor_list_worlds_latencyp(95)<300
hathor_get_world_latencyp(95)<250
hathor_export_world_latencyp(95)<2000p(99)<4000
errorsrate<0.05

isis

ENQUEUE_THRESHOLDS#
http_req_failedrate<0.05
enqueue_success_raterate>0.95
enqueue_submit_latencyp(95)<1200p(99)<2500
errorsrate<0.05
WORKER_THRESHOLDS#
http_req_failedrate<0.05
status_poll_latencyp(95)<700p(99)<1500
queue_position_latencyp(95)<900
worker_running_raterate>0.05
worker_terminal_raterate>0.10
errorsrate<0.05

sophia

GRAPH_THRESHOLDS#
http_req_failedrate<0.02
resolve_success_raterate>0.98
subgraph_success_raterate>0.95
sophia_resolve_latencyp(95)<300p(99)<600
sophia_subgraph_latencyp(95)<600p(99)<1200
errorsrate<0.05
SEARCH_THRESHOLDS#
http_req_failedrate<0.02
search_success_raterate>0.98
ask_success_raterate>0.95
sophia_search_latencyp(95)<500p(99)<1000
sophia_ask_latencyp(95)<800p(99)<1500
errorsrate<0.05

veritas

STANDARD_THRESHOLDS#
http_req_durationp(95)<500
http_req_failedrate<0.01
errorsrate<0.05
api_latencyp(95)<300
article_latencyp(95)<250
search_latencyp(95)<500
auth_latencyp(95)<1000
fact_check_latencyp(95)<3000
source_latencyp(95)<300
subscription_latencyp(95)<500
payment_latencyp(95)<2000
taxonomy_latencyp(95)<200
feed_latencyp(95)<400
cdn_latencyp(95)<200
db_query_latencyp(95)<300
rate_limitedrate<0.02
STRESS_THRESHOLDS#
http_req_durationp(99)<3000
http_req_failedrate<0.10
errorsrate<0.10
article_latencyp(95)<1000
search_latencyp(95)<2000
auth_latencyp(95)<3000
rate_limitedrate<0.15

yemaya

ASSET_THRESHOLDS#
http_req_failedrate<0.02
asset_register_success_raterate>0.98
asset_commit_success_raterate>0.98
yemaya_asset_register_latencyp(95)<400
yemaya_asset_commit_latencyp(95)<600
yemaya_asset_upload_total_latencyp(95)<2000p(99)<4000
errorsrate<0.03
COLLAB_THRESHOLDS#
http_req_failedrate<0.02
presence_success_raterate>0.99
yemaya_presence_latencyp(95)<250p(99)<500
errorsrate<0.03
EDIT_THRESHOLDS#
http_req_failedrate<0.02
project_edit_success_raterate>0.99
project_create_success_raterate>0.98
yemaya_studio_edit_latencyp(95)<150p(99)<300
yemaya_project_create_latencyp(95)<400p(99)<800
errorsrate<0.03