I’ve been running InsightCloudSec (formerly DivvyCloud) for about nine months across a consolidated payer account with roughly 200 linked AWS member accounts. The initial promise was solid: a single pane of glass for cloud security posture, compliance, and asset inventory. The reality, especially over the last quarter as we’ve scaled, is that the platform has become borderline unusable for anything resembling real-time analysis. The web UI is a laggy mess, and API calls for basic inventory data routinely time out.
My team’s primary use case is generating a daily aggregated view of specific resource types (S3 buckets, IAM roles, EC2 instances with public IPs) for our FinOps and SecOps dashboards. What used to take a minute or two now consistently fails or takes upwards of 15-20 minutes. I’ve engaged with their support, who of course suggested the usual: scaling up the collector appliances, ensuring we’re on the latest version, and pruning our collection scope. We’ve done all that. The bottleneck appears to be in their data aggregation layer when dealing with a high cardinality of resources across many accounts. For example, a simple API call to list all S3 buckets with tagging data shouldn’t require client-side pagination logic that has to wait 30 seconds between each page fetch.
Here’s a sanitized snippet of the script we use, which now fails regularly due to HTTP 504s from their gateway:
```
import requests
from requests.auth import HTTPBasicAuth
BASE_URL = "https://our-instance.insightcloudsec.com/api/v2/public"
ENDPOINT = "/inventory/search"
QUERY = {
"query": {
"resource_type": ["aws::s3::bucket"],
"fields": ["resource_id", "account_name", "tags"]
},
"paginate": True,
"page_size": 100
}
response = requests.post(
f"{BASE_URL}{ENDPOINT}",
auth=HTTPBasicAuth(API_KEY, ""),
json=QUERY,
timeout=300
)
# This now often times out before returning the first page
```
We’re running their recommended “large” collector setup in our primary region. The collectors themselves don’t seem overly taxed—CPU and memory are within bounds. The slowness feels architectural, like their backend is doing some monstrous join across every table in their schema for each request, and it doesn’t scale linearly. Has anyone else hit this wall with large, multi-account AWS environments? Did you find any tuning parameters they don’t advertise, or did you have to move to a different product for inventory and real-time querying? I’m starting to think the marketing around “real-time cloud security” is at odds with the reality of their data store performance at scale.
-- Cam
Trust but verify.