Skip to main content

Operational Tuning

Adjust application behavior in realtime without deploying code.

Rate Limits​

Store rate limit configurations:

rate-limits
{
  "api-requests-per-minute": 100,
  "api-requests-per-hour": 5000,
  "max-concurrent-connections": 50,
  "max-payload-size-mb": 10
}

Use in your API:

import { createReplaneClient } from 'replane-sdk';

const client = createReplaneClient({
  apiKey: process.env.REPLANE_API_KEY,
  baseUrl: process.env.REPLANE_URL,
});

const limits = await client.watchConfigValue('rate-limits');

// In your rate limiter
function getRateLimit() {
  return limits.get()['api-requests-per-minute'];
}

When you need to increase limits during a traffic spike:

  1. Update the config in the Replane UI
  2. Changes propagate instantly to all app instances
  3. No restart required

Cache TTLs​

Tune cache behavior dynamically:

cache-config
{
  "user-profile-ttl-seconds": 300,
  "product-catalog-ttl-seconds": 3600,
  "homepage-ttl-seconds": 60
}

Implementation:

const cacheConfig = await client.watchConfigValue('cache-config');

async function getCachedUserProfile(userId) {
  const ttl = cacheConfig.get()['user-profile-ttl-seconds'];
  return cache.get(`user:${userId}`, { ttl });
}

Batch Sizes​

Control background job batch sizes:

job-config
{
  "email-batch-size": 100,
  "webhook-batch-size": 50,
  "export-batch-size": 1000
}

Use in your worker:

const jobConfig = await client.watchConfigValue('job-config');

async function processEmails() {
  const batchSize = jobConfig.get()['email-batch-size'];
  const emails = await getEmailQueue(batchSize);
  // Process batch...
}

Timeouts & Retries​

Store timeout and retry policies:

resilience-config
{
  "api-timeout-ms": 5000,
  "database-timeout-ms": 3000,
  "max-retries": 3,
  "retry-delay-ms": 1000
}

Feature Thresholds​

Control when features trigger:

thresholds
{
  "free-tier-max-items": 100,
  "premium-tier-max-items": 10000,
  "bulk-import-min-items": 50,
  "large-file-warning-mb": 50
}

Circuit Breaker Settings​

Manage circuit breaker behavior:

circuit-breaker
{
  "failure-threshold": 5,
  "timeout-ms": 10000,
  "reset-timeout-ms": 30000
}

JSON Schema for Safety​

Prevent invalid configurations:

Schema for rate-limits
{
  "type": "object",
  "properties": {
    "api-requests-per-minute": {
      "type": "integer",
      "minimum": 1,
      "maximum": 10000
    },
    "max-concurrent-connections": {
      "type": "integer",
      "minimum": 1,
      "maximum": 1000
    }
  },
  "required": ["api-requests-per-minute"],
  "additionalProperties": false
}

This ensures values stay within safe ranges.

Best Practices​

Start Conservative​

Begin with safe, conservative values:

{
  "api-requests-per-minute": 50,  // Start low
  "max-concurrent-connections": 20
}

Increase gradually based on monitoring.

Monitor Impact​

After changing a config:

  • Watch error rates
  • Monitor latency
  • Check resource usage

Document Units​

Be explicit about units in config names:

// ❌ Ambiguous
"timeout": 5

// ✅ Clear
"timeout-ms": 5000
"cache-ttl-seconds": 300

Use Watchers for Hot Paths​

For frequently accessed configs, use watchConfigValue:

// ❌ Fetches on every request
async function rateLimit(req) {
  const limits = await client.getConfigValue('rate-limits');
  // ...
}

// ✅ Cached in memory, updated in realtime
const limits = await client.watchConfigValue('rate-limits');

function rateLimit(req) {
  const rpm = limits.get()['api-requests-per-minute'];
  // ...
}

Rollback Plan​

If a config change causes issues:

  1. Go to Replane UI
  2. Click "Version History"
  3. Select the previous version
  4. Click "Rollback"

Changes propagate instantly.

Incident Response Example​

Scenario: API is overloaded during traffic spike

  1. Open Replane UI
  2. Navigate to rate-limits config
  3. Increase api-requests-per-minute from 100 to 200
  4. Save
  5. All app instances update within seconds
  6. Monitor metrics
  7. Adjust further if needed

No deploy, no restart, no downtime.

Next Steps​