Affordability

The trial comes first. New DevOps Agent customers get a two-month free trial with up to 20 hours of investigations per trial month. No check reliably says whether a given account is still in it, so say any estimate below applies to usage beyond the trial allowance, or after the trial ends.

What you need, and what to do without it

You need AWS credentials in the environment. You do not need the user signed into AWS Startup Advisor — this runs on whatever aws configure, AWS_PROFILE, or SSO session is already there.

The answer does not require the data. Even with no numbers you can say: “Each investigation is about $3–4 after the trial, and it’s easy to leave running. Point it at production only, and check back after a week.” That is still more useful than a price list.

The arithmetic

$4 is the top of the per-investigation range and the right number to plan against.

Estimating investigations per week. Best evidence first:

1. Operational alerts already firing in Startup Advisor — lambda-throttled, asg-health-check, route53-failover. These are real, current, and countable.
2. Ask, in their terms: “Roughly how often does something break badly enough that you drop what you’re doing?” Founders answer that accurately; they do not answer “what’s your incident rate.”
3. Default to 3 a week if there is nothing to go on. Low enough not to alarm, high enough to be honest.

Saying the result

Give the number plainly — “after the trial, at about three investigations a week (about $12), your $200 in credits would last roughly four months, unless they expire sooner or other AWS usage draws on them” — and say what happens when credits run out (usage bills to their payment method). If it is short, they may want the lower-cost options below; mention them, but do not recommend against setup. A founder who has just lost a day to an outage may reasonably decide four dollars is cheap.

Never assume a trial is active, and never invent a balance or runway the data does not support.

Lower-cost options, if they want them

In descending order of value:

1. Release-readiness review, which is free during preview. The strongest answer by some distance. It costs nothing today, works with zero production traffic, and prevents a share of the incidents they cannot currently afford to investigate — same product, no cost, real value.

2. CloudWatch alarms. If Startup Advisor is reporting scalability.no-cloudwatch-alarms, that is the gap to close first and it is nearly free — the AWS free tier covers ten alarms. An alarm that pages them at 2 a.m. beats an agent they cannot afford to run.

3. Structured logging and log retention. cost.cloudwatch-no-retention often shows up alongside. Being able to search logs is the prerequisite for diagnosing anything by hand.

4. The resiliency baseline and architecture assessment prompts, already in Startup Advisor and free.

5. Come back at the first incident they cannot diagnose in thirty minutes. Concrete, and it is their own threshold rather than a sales trigger.

Timing traps worth flagging

• Activate credits expiring inside the two-month trial. If credits lapse before the trial ends, free becomes paid and the balance vanishes at roughly the same moment. Check endDate against the trial and say so if they collide.
• A launch is coming. Product Hunt, demo day, a big customer going live — incident rate and cost both spike exactly when the founder is least able to think about it. Worth raising before the launch, not during.
• The first 24 hours after connecting are the weakest. The agent is still learning topology. Investigations get sharper; the first one is not the one to judge it on.