MorpheusLens

โ— CW Live (region-specific) ยท โ— OS Manual (all regions combined)
3/3 operational regions reporting Click a region to investigate

Service Velocity

Are we completing work faster than new work is becoming non-progressing?
โš  Data source: OpenSearch (manually fetched) ยท Last updated:
Compares validations completing per day (green) vs new validations becoming non-progressing per day (red). Non-progressing = created more than 7 days ago and still in non-terminal state. When red exceeds green consistently, problems are accumulating.
Completions/day vs New Non-Progressing/day (60-day view) โฑ OpenSearch

Business Proof

Evidence that Morpheus is delivering value
โš  Data source: OpenSearch (manually fetched) ยท Last updated:

Live Operational Metrics

CloudWatch (1-week window, 5-minute data points)
โ— Data source: CloudWatch (live) ยท Auto-refreshes every 60s ยท Requires Midway auth
Live metrics from AWS CloudWatch. Each widget has interactive time controls. Red dashed lines = alarm thresholds from CDK.

Validation Age Distribution

How long have non-terminal validations been in the system?
โš  Data source: OpenSearch (manually fetched) ยท Last updated:
Validations in non-terminal status (RUNNING, CREATED, POST_CHECK_IN_PROGRESS) grouped by age. Validations naturally take hours to days โ€” the P90 completion time is ~3.8 days. The 30d+ buckets indicate work that likely needs intervention.
Non-Terminal Validations by Age (Production Only) โฑ OpenSearch
โ–ถ Oldest Non-Terminal Validations (Detail Table) Click to expand โ€” individual validations sorted by age

Orchestrator Queue & Data Pipeline

SQS queue connecting the API to the Orchestrator โ€” is work being picked up or accumulating?
โ— Data source: CloudWatch (live) ยท SQS metrics for the ValidationEventQueue ยท Auto-refreshes every 60s
โ–ถ Check Failures โ€” By Type & Name Click to expand โ€” which checks are failing and their specific test names
โ–ถ Device Health Click to expand โ€” device check pass/fail rates and top failing hardware models
โ–ถ Infrastructure Health Click to expand โ€” ECS CPU/Memory, Watchdog Lambda health, deployment context
โ— Data source: CloudWatch (live) ยท ECS ContainerInsights + Lambda metrics

Open Validations Analysis

Deep dive into validations that have not reached a terminal state
โš  Data source: OpenSearch (manually fetched) ยท Last updated:
Open validations are those in non-terminal status (CREATED, RUNNING, POST_CHECK_IN_PROGRESS) that have not yet reached SUCCEEDED, FAILED, or CANCELLED. This analysis helps identify validations that may need intervention โ€” either because the orchestrator lost track of them, a downstream system stopped responding, or they were created and never started.
Status Breakdown โฑ OpenSearch
Age Distribution โ€” Are These Truly Dead?
Every validation has a lastUpdatedTime that updates whenever any action occurs. If this hasn't changed, nothing has happened โ€” no orchestrator activity, no check execution, no user interaction.

Context: Validation cycle time P90 is ~3.8 days. Validations in the 1-7d bucket are likely actively progressing. The 30d+ buckets represent work that will almost certainly never complete without manual intervention.

Open Validations by Workflow Type
Monthly Trend โ€” New Open Validations Created
โ–ถ Oldest Open Validations (Full Detail) Click to expand โ€” individual validations sorted oldest first with workflow and link count
Recommendations

Morpheus System Architecture

How requests flow through the system
Morpheus validates network links and devices during datacenter scaling. When a scaling team creates a validation, it flows through API โ†’ Orchestrator โ†’ Check Workers โ†’ results stored back.