Troubleshooting — Performance & latency¶

When the Console is sluggish, when API calls feel slow, when objects take forever to download — work through this page. Most performance issues come down to: where is the user, where is the resource, and what's on the path between them.

Mental model: where's the latency budget going¶

graph LR
    User[User device]
    ISP[Customer ISP]
    BDIX((BDIX))
    Edge[Cloud Digit edge]
    Region[Region: bd-dha-1 / bd-ctg-1 / bd-syl-1]
    Service[Service backend]

    User --> ISP --> BDIX --> Edge --> Region --> Service

Typical budgets for a domestic Bangladeshi user:

If you're seeing 200+ ms, something's wrong on the path. The diagnostic flow:

graph TD
    A[Console feels slow] --> B[Open DevTools Network tab]
    B --> C{First-byte time}
    C -->|"< 100 ms"| D[Frontend issue]
    C -->|"100-500 ms"| E[Edge / region issue]
    C -->|"> 500 ms"| F[Routing or service issue]

    D --> D1[CPU / extensions / dev console open]
    E --> E1[Region selection / CDN warm-up]
    F --> F1[International transit / service incident]

Console UI feels slow¶

API calls are slow¶

For API consumers (CLI, Terraform, your own apps):

Measuring from your side¶

```bash

Time a single API call end-to-end¶

curl -w "@curl-format.txt" -o /dev/null -s -H "Authorization: Bearer $CD_API_TOKEN" \ https://api.bd-dha-1.clouddigit.ai/v1/compute/servers

Where curl-format.txt is:¶

time_namelookup: %{time_namelookup}\n¶

time_connect: %{time_connect}\n¶

time_appconnect: %{time_appconnect}\n¶

time_pretransfer: %{time_pretransfer}\n¶

time_redirect: %{time_redirect}\n¶

time_starttransfer:%{time_starttransfer}\n¶

time_total: %{time_total}\n¶

```

Compare time_starttransfer to the typical-budget table above. If your overall is > 100 ms while staying domestic, escalate.

Object Storage uploads/downloads¶

VM / Kubernetes networking¶

Database¶

For deeper DB ops, see Managed DBA.

"From inside Bangladesh it's fast, from outside it's slow"¶

Cloud Digit is sovereign-resident — services are in BD. Reaching them from outside BD adds international transit latency. Patterns:

This is inherent — no fix on Cloud Digit's side. Mitigations:

• Front public-facing services with CDN for international reads
• For interactive admin work, accept the latency; for bulk transfers, schedule overnight
• Cache aggressively in app layer

When to open a ticket¶

Open a ticket when:

• Performance is materially worse than the budgets above and reproducible
• You've ruled out client-side issues (close DevTools, test in incognito, test from another machine)
• Multiple users / multiple resources are affected
• The performance has changed recently without changes on your end

Include:

• Trace IDs of slow API calls (X-Cd-Trace-Id response header)
• Timestamps + region(s) affected
• curl -w output for representative calls
• Network path (where users / clients are; ISP names)
• Whether the issue is constant or intermittent (with intermittent pattern if any)

Related¶

• Browsers, sessions & SPA quirks
• Troubleshoot — Browser & cache
• Networking — VPC + security-group setup
• BDIX Peering Direct Connect — for sub-ms domestic
• CDN — for international fan-out
• Status — platform-wide latency reports
• Managed Monitoring / SIEM — for application-side observability