You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
'Looks like the fixability endpoint in getsentry/seer could use some more granular span and profiling instrumentation. Uptime monitoring would also be helpful since it seems to be intermittently unavailable.',
82
-
products: ['tracing','profiling','uptime'],
80
+
purpose:
81
+
'Improve observability for the fixability endpoint to better track intermittent failures when called from Seer Scanner.',
82
+
observability_requests: [
83
+
{
84
+
description:
85
+
'Add detailed tracing spans around the ML model inference pipeline to track processing time and identify bottlenecks in the fixability assessment workflow.',
86
+
instrument_type: 'tracing',
87
+
location:
88
+
'getsentry/seer/src/seer/automation/fixpoint/api.py in the process_fixability_request function',
89
+
},
90
+
{
91
+
description:
92
+
'Implement profiling to identify memory usage patterns and CPU hotspots during model loading and inference phases.',
93
+
instrument_type: 'profiling',
94
+
location:
95
+
'getsentry/seer/src/seer/automation/models/ module, specifically in model initialization and prediction methods',
96
+
},
97
+
{
98
+
description:
99
+
'Add structured logging with correlation IDs to track requests through the entire pipeline and capture error context.',
100
+
instrument_type: 'logging',
101
+
location:
102
+
'getsentry/seer/src/seer/automation/fixpoint/api.py throughout the request lifecycle',
103
+
},
104
+
{
105
+
description:
106
+
'Tag requests with metadata like model version, request size, and processing mode to enable better filtering and analysis.',
107
+
instrument_type: 'tagging',
108
+
location:
109
+
'getsentry/seer/src/seer/automation/fixpoint/api.py at the entry point of each request handler',
0 commit comments