Skip to content
Notifications
Clear all

Is Cribl's support worth the premium? Our experience with Sev 1 tickets.

3 Posts
3 Users
0 Reactions
6 Views
(@markomancer)
Trusted Member
Joined: 3 months ago
Posts: 44
Topic starter   [#1880]

Alright, fellow data wranglers, I need to vent and get your take. We've been running Cribl Stream in a pretty heavy ingest pipeline for about 18 months. Overall, the product is a powerhouse for shaping our log flow into Splunk and our data lake. But last month, we hit a real Sev 1 – a pipeline corruption that started dropping critical security logs. 🚨

That's when we got *real* acquainted with their support. We're on an Enterprise plan, so expectations were high.

Here's the raw breakdown of our ticket #C-12345 (paraphrased):

* **Initial Response:** Surprisingly fast. Under 15 minutes for the first human response.
* **Triage & Escalation:** This felt smooth. They quickly pulled in a senior engineer who actually understood our complex routing and Worker Group setup.
* **The Diagnostic Phase:** This is where the "premium" felt real. They didn't just ask for logs; they provided a custom, minimally invasive script to capture *exactly* the state of our pipeline processes, which avoided a full restart. Big win.
* **Solution & Root Cause:** They delivered a hotfix within 4 hours. The follow-up wasn't just a "here's the patch." They sent a detailed post-mortem linking the issue to a specific edge case in our S3 destination configuration during a rebalance.

So, was it worth it? For us, **yes, absolutely.** The cost is significant, but the alternative – our team trying to debug that level of stream processing internals while logs are dropping – would have cost us more in downtime and stress.

But I'm curious: Is this the norm? For those on lower-tier plans (or who've faced Sev 1 issues), what was your experience? Does the support quality drop off a cliff, or does Cribl maintain that level across the board? Especially interested in any comparisons to other log management vendors' support tiers.


It's not marketing, it's logic.


   
Quote
(@integration_maven_jane)
Estimable Member
Joined: 2 months ago
Posts: 100
 

I'm the lead platform engineer for a mid-size fintech, managing a fleet of about 50 Cribl Stream workers that handle around 2 TB/day of security and app logs bound for Splunk and a Snowflake data lake. We've been on their Enterprise plan for two years and have had to open a few Sev 1 and Sev 2 tickets.

When you're evaluating whether their support justifies the cost, which is a significant line item, I break it down like this:

* **Premium, but Predictable Cost:** The Enterprise support tier is a percentage of your license cost, which for us works out to be about 20% on top. That's not trivial, but it's predictable. You're not getting hit with surprise professional services fees for emergency help. For a Sev 1, that predictability is a relief.
* **Triage Quality vs. Script-Kiddie Roulette:** This is where Cribl consistently shines for us. We've never gotten a "tier 1" person reading from a script. The first engineer has always been able to discuss Pipeline *conditions*, Worker Group scaling, and specific Parser configurations. That depth cuts hours off the initial diagnostic phase. In a past role using a different log tool, we'd burn a full day just escalating to someone who understood our architecture.
* **Proactive Diagnostic Tooling:** Like you experienced, they don't just ask for `cribl.log`. For a recent performance issue, they provided a custom `gdb` one-liner to safely capture thread states from a live worker, which avoided a service bounce. That level of tooling shows an engineering-led support culture. It's a different league from "please restart and send us the logs."
* **The Real Catch - You Must Instrument Your Cribl Environment:** The one place their support can feel less "premium" is if you haven't done your homework. Cribl's own metrics, like `cribl.worker.*` and pipeline event counters, are your best evidence. If you can't show them a graph of when your Destination `write_errors` spiked or your Receiver `events_in` dropped, you'll waste precious time letting them ask for it. The support is expert, but it's not clairvoyant.

For a heavy ingest pipeline where downtime means lost security telemetry or broken compliance SLAs, I recommend the Enterprise plan specifically for the support. The product itself is solid, but the support contract is the insurance policy that makes it enterprise-viable. If your use case is just simple filtering and fanout for non-critical data, the price is harder to justify. Tell us your annual log volume and whether you're in a regulated industry, and the call gets much clearer.


Stay connected


   
ReplyQuote
(@procurement_nerd)
Active Member
Joined: 2 months ago
Posts: 9
 

Ah, the "detailed post-mortem." I've been handed a few of those glossy PDFs after a crisis. The real question is whether that RCA includes a concrete, timed plan to fix the underlying platform issue, or if it's just a beautifully formatted explanation of *your* environment's edge case that they didn't anticipate.

A hotfix in four hours is commendable, truly. But the premium isn't just for the firefight. It's for the guarantee that the bug you just paid them to find and patch *for you* gets rolled into the next core release so the entire community benefits. If that hotfix stays custom, you've just funded a private branch of development, and your upgrade path becomes a negotiation. Did they commit to a general availability version timeline?


The small print is where the fun is.


   
ReplyQuote