{
  "id": 2117534,
  "title": "The $11k cloud bill was mostly the tools you added to watch the cloud bill",
  "url": "https://urgent.news/2026/08/20/the-11k-cloud-bill-was-mostly-the-tools-you-added-to-watch-the-cloud",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-08-20T10:38:23.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/adioof/the-11k-cloud-bill-was-mostly-the-tools-you-added-to-watch-the-cloud-bill-3oec"
  },
  "original_language": "en",
  "account": "A startup with six engineers found their monthly cloud bill to be $11,847. The largest portion of this expense came from monitoring tools rather than the application itself. This pattern is common, as seen in a Medium post from August 13, 2026, where a team reduced their bill from $11,847 per month to $3,812 by cutting overhead from four services that could run on two servers.\n\nSoftware engineer Devrim Ozcay detailed a case where his AWS bill was $1,850, but Datadog's bill reached $3,200 and even peaked at $12,000 due to container auto-scaling. Similarly, a nine-microservice and Lambda stack had AWS compute costs of $18,432 and Datadog costs of $24,773. Another developer forum thread discussed a firm spending $52,000 on AWS but paying $97,000 for observability, with $47,000 going to Datadog, $38,000 to Splunk, and $12,000 to Sentry.\n\nSidecar containers used in service meshes can consume significant resources, with one expert noting they use 50-100MB of RAM and 10-50m CPU per pod, adding 2-5ms of latency per network hop. Cloud architect Khimananda Oli expanded on this, stating that sidecar containers can consume up to 200MB of RAM each and introduce up to 15ms of latency, potentially burning a full gigabyte of RAM for five microservices.\n\nDespite the high costs, observability is important. However, many organizations are realizing that their architecture may be to blame. An estimated 42% of microservices adopters have consolidated services back into larger units due to complexity in debugging and high costs of network latency. The bill is not a compute problem but an architectural one. To address this, organizations should reduce the surface area and then focus on instrumenting the remaining parts. The key takeaway is that observability should not be more expensive than the entities being observed; instead, reducing the number of entities being observed is often the solution.",
  "summary": "A startup with only six engineers calculated the cost of their monthly cloud bill to be $11,847. The majority of the data generated was not from the application itself, but from the monitoring tools that were observing the application. ## The recursive joke nobody's laughing at We often see the same pattern in the industry. You create a small app. Next, you attach the equipment needed to monitor…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}