Best offer on record: Datadog Free Plan (Infrastructure monitoring). 5 verified findings, checked 2026-10-22.
Offers
Datadog Free Plan (Infrastructure monitoring) high confidence
Free monitoring plan at $0: up to 5 hosts, 1-day metric retention, 5 containers per host (FAQ says '5 or 10 containers free for each host license'), unlimited user accounts, 'core collection and visualization features', 1,000+ integrations. The current pricing table does NOT show custom-metric/custom-event allotments in the Free column; its FAQ attributes '100 custom metrics/500 custom events per host' to Pro and '200/1000' to Enterprise — so the previously advertised 100-metric/500-event free allotment is no longer explicitly granted on the Free tier as displayed today. Paid reference points: Infrastructure Pro $15/host/mo (annual, 15-month retention), Enterprise $23/host/mo.
Eligibility
Open to any new Datadog account holder; no .edu email, no startup proof, no credit card stated for the free plan; no region restriction stated. Data region chosen at signup cannot later be migrated.
How to claim
Sign up at app.datadoghq.com/register (no credit card required) -> choose data region -> the 5-host Free plan applies automatically to the account; no application or referral needed.
Expiry
Indefinite; no expiration or renewal. 1-day metric retention means data older than 1 day is purged; plan limits (5 hosts) are fixed.
Gotchas
1-day retention vs 15 months on paid plans; container allotment (5 or 10 per host) is a soft cap; over 5 hosts or over 10 containers/host requires a paid plan; the 100 custom metrics/500 custom events per-host allotments appear tied to paid Pro/Enterprise in today's pricing FAQ, so verify current free-tier custom-metric behavior in-product. 'Unlimited alerts' and 'unlimited users' are not explicitly printed in the Free column on today's page (they appear under Pro/Enterprise rows).
Agent Observability (LLM Observability) Free Tier high confidence
Per today's pricing page: Free tier gives 'Up to 40K LLM spans / month', '15-day trace retention', 'Unlimited context & evals', 'Full feature access', positioned 'For individuals and small teams', and the FAQ states 'up to 40K LLM spans per month with 15-day retention ... no credit card required.' Paid contrast: Pro 'Starting At $160 per month' includes 100K LLM spans/month; overage 'Additional spans: $3.50 per 10K LLM spans'.
Eligibility
Any Datadog account holder; individuals and small teams; no .edu or startup proof; no credit card required; no region limit stated.
How to claim
Included with a standard Datadog account (sign up free, no CC) -> connect LLM/agent telemetry via OpenTelemetry OpenInference integrations or supported LLM SDK integrations -> spans are tracked up to 40K/month with no application.
Expiry
40K LLM-span allowance resets monthly; 15-day rolling trace retention; no end date. Exceeding 40K spans on the free tier pushes you to the paid Pro tier pricing.
Gotchas
Renamed from LLM Observability to Agent Observability; 'full feature access' is on the free tier but the product was recently GA-transitioned, so some features may be in preview; 15-day retention is short vs 15-month metric retention on paid infra plans; span overages are not billed on free — they are the trigger to upgrade; egress/scanning costs (e.g., AI Guard at $0.80/1M tokens scanned) are separate paid line items if enabled.
14-day Free Trial (metrics, APM/traces, logs, Synthetics) high confidence
Official trial page: '14 days of unlimited monitoring' covering 'metrics, traces, and logs', log search/analysis, APM request tracing, and Synthetic-like monitoring of 'applications and API endpoints via simulated user requests', plus anomaly/outlier/forecast alerts. 'No credit card required' is stated on the page. 'Unlimited' during trial means no host limit for the trial window; at trial end the account drops to plan-based limits (free plan for infra).
Eligibility
Any new Datadog user; no .edu email or startup proof required; no credit card; region choice at signup is binding ('You can't migrate data between regions').
How to claim
Click any 'Start Free Trial' CTA (or datadoghq.com/trial) -> create account with no CC -> pick data region -> monitoring is live immediately for 14 days; no application, no sales call.
Expiry
14 days from signup; not renewable as a trial (one trial per account in practice). After it ends, only the free-plan limits (e.g., 5 hosts) and the 40K LLM-span tier apply; paid products require a plan.
Gotchas
Trial-vs-free confusion: 'unlimited' applies only during the 14 days; post-trial you keep the free plan, not unlimited monitoring. No post-trial behavior is stated on the page (data retention/downgrade specifics not published). Region cannot be changed later. Trial does not include all products (e.g., Cloud SIEM/Security add-ons have separate trial offers).
Datadog for Startups (DDFS) — up to $100K credits high confidence
Official startup page: 'up to a year of free access to the full Datadog platform — up to $100k in credits', hero states 'a year of free Datadog Pro'. Coverage: complete GA product suite (Observability: infrastructure/APM/logs/data; Security: code/cloud/SIEM; Digital Experience; Software Delivery; Service Management; AI modules) with 'No watered-down tiers, no hidden limits'; preview products 'if space permits'. Extras: optional Technical Engagement Manager, documentation access, customer-success rep after several months.
Eligibility
(1) Funding stage 'Series A or earlier' (the page's cap is funding-stage based, no explicit employee/revenue figures published); (2) 'New to Datadog' — no current/prior Datadog customers and no prior DDFS credits; (3) 'Referred by one of our partners' (VCs, hyperscalers, accelerators) — applications without a partner referral are not accepted. No .edu component.
How to claim
Get a referral from a partner (VC/hyperscaler/accelerator) -> submit the application at datadoghq.com/startups describing your team -> start a Datadog trial while it is reviewed -> acceptance notified in ~3-5 business days -> activate within 30 days after acceptance or credits are forfeited.
Expiry
Credits valid up to 1 year from acceptance/activation; activation window is 30 days post-acceptance; inactive accounts (3 months without usage) get a 30-day warning before removal; unused credits do not roll over.
Gotchas
Partner-referral-only — no direct application path, which excludes bootstrapped/unaffiliated startups; 'up to $100k' is a ceiling, actual awards vary; credits apply to services, not products you can't resell; starting a trial during review is required by the process; three-month inactivity rule can void remaining credits.
Student/education license (status: no public offer found as of 2026-10-22) medium confidence
No public Datadog education program, student license, or .edu-based free tier was found on the official site today: datadoghq.com/education/ returns 404; /solutions/education/ is a sales page aimed at selling Datadog TO institutions (Harvard Medical School, University of Chicago, Coursera as customers), with no free seat grants; /learn/ offers a free 'Learning Center' (hands-on course platform for all users, free of charge) but no student/educator account or seat grants, no .edu email requirement, and no classroom environment program is advertised (the older 'Datadog Classroom' URL structure is no longer live).
Eligibility
Not applicable — no public application process. The general 14-day trial and 5-host free plan are open to students/educators like anyone else (no CC required for trial; no CC stated for free plan), and free learning courses at /learn/ are open to all users.
How to claim
N/A — no education application exists. Nearest path: use the general free trial (no CC) + 5-host free plan for coursework; use free /learn/ courses for certification-style study.
Expiry
N/A for the (absent) program; free plan and trial terms as in their own findings.
Gotchas
Students have no dedicated .edu offer or classroom org grants discoverable today; do not assume a 'Datadog Classroom' free lab still exists; confirm with Datadog sales/academics if a class agreement is needed — any such access would be custom, not public.