Customer Knowledge Base
Customer Knowledge Base
Breadcrumbs

HCMS: Performance Degradation & Stale Data During High Load

1. Overview

This KB explains why the Headless CMS (HCMS) may show degraded performance or temporarily return outdated data during periods of elevated load. It separates the two distinct types of load that can affect HCMS and describes how each impacts system behavior, how to diagnose it, and what can be done to mitigate it. 

2. Symptoms

  • Slow API responses

  • Increased response‑time variance

  • Intermittent unavailability

  • Occasional 502/503 errors (depending on hosting setup)

  • Returned data appearing older than expected

  • Delays before updates become visible in HCMS

3. Common Causes

  • High read load on HCMS components

  • High write load on the censhare Server

  • Large or continuous data operations (imports, migrations, sync processes)

  • Parallel or unthrottled integration calls

  • Application‑level patterns generating excessive queries

  • Peaks in usage combined with background workloads

4. Diagnostics

4.1 Customer/Partner

  • Check whether the issue coincided with:

    • High‑frequency or complex read operations

    • Large write operations (migrations, imports, sync jobs)

    • Automated or parallel integration calls

  • Verify whether updated data appears correctly in censhare Core

  • Retry the same request after a short interval and note whether performance improves

  • Review integration patterns for unnecessary polling or excessive parallelism

4.2 Infra/DevOps

  • Review system metrics (CPU, memory, I/O) for HCMS satellites

  • Review metrics on the censhare Server side to identify write‑load saturation

  • Check for restarts or slowdowns of HCMS components

  • Validate whether traffic patterns indicate unusual spikes or potential attack behavior

4.3 censhare Support

  • Validate whether sync delays occurred during the issue timeframe

  • Analyze whether load patterns indicate excessive reads or excessive writes

  • Confirm whether the behavior correlates with known system limits or ongoing heavy operations

5. Mitigation

High Load on HCMS (Read Load)

Occurs when HCMS receives a large number of complex or frequent read requests.

Effects:

  • Slow API responses

  • Longer response times

  • Possible temporary unresponsiveness

Mitigation:

  • Add additional HCMS satellites to distribute load

  • Increase CPU and memory (helps, but to a limited extent)

  • Review and optimize the consuming application’s query patterns

  • Investigate potential abnormal traffic or (D)DOS activity; if identified, implement a WAF or external rate limiting (not part of HCMS)

High Load on the Server (Write Load)

Occurs when censhare Server is processing excessive or large write operations.

Effects:

  • Delayed sync to HCMS

  • HCMS returning outdated data until sync catches up

  • Slow or timed‑out write operations (timeouts often around 60 seconds by default)

Mitigation:

  • Reduce or batch large write operations

  • Distribute heavy workloads over time

  • Review application‑level write behavior and adjust strategy

  • Evaluate infrastructure to ensure adequate capacity for sustained write throughput

Combined Load Situations

Both read and write load conditions can occur at the same time. In such cases, both sets of mitigation steps may be required.

6. Best Practices

  • Batch heavy operations and avoid running large migrations during peak activity

  • Implement throttling or rate limiting in integrations

  • Avoid unnecessary polling and use incremental data update strategies

  • Add capacity during known heavy use periods

  • Ensure monitoring is in place for request volume, load patterns, and sync lag

7. When to Contact Support

Contact censhare Support if:

  • Performance issues persist after reducing load

  • Stale data continues even after sync should have completed

  • Response times degrade unpredictably and without identifiable workload patterns

8. Information Required

To accelerate analysis, provide:

  • Timeframe when the issues occurred

  • Whether any large operations (imports, sync runs, migration steps) were active

  • Affected endpoints or request types

  • Example requests and responses showing stale or slow behavior

  • Whether updated data appears in censhare Core

  • Whether the load was primarily read‑heavy, write‑heavy, or both