Compare / Elastic Serverless
WarpSearch vs Elastic Cloud Serverless
Elastic's serverless Elasticsearch. Search projects bill compute (VCU-hours) plus retention. Observability and Security projects bill per GB ingested.
The scoreboard
Price
Cheaper on 2 of 4 workloads
Elastic Serverless is cheaper for: Mid-size logs, Large logs.
Monthly, list prices. Bars are drawn to the middle of a range.
Performance
Likely faster
Likely faster, not like for like
Elastic publishes p99 of 22 to 67 ms on its own Rally runs. Our medians are 2 to 16 ms on 2 to 4 CPUs. Median against p99 on different hardware is indicative only; we will publish p99 on the same datasets.
How they bill
| WarpSearch | Elastic Serverless | |
|---|---|---|
| Billing unit | Per GB stored, per GB written, per request | Search: VCU-hours ($0.14 ingest, $0.09 search) plus $0.047/GB retention. Logs: $0.07 to $0.09/GB ingest plus $0.017 to $0.019/GB retention |
| Scale to zero | Yes, after 15 minutes idle | Ingest scales to zero. Search keeps a baseline, billed at a reduced idle rate |
| API | Elasticsearch compatible | Native Elasticsearch |
| Hybrid and rank fusion | Keyword and vector in one query; RRF coming | RRF and linear retrievers |
| Aggregations | Common aggregations | Full set |
Price
Monthly cost for four common workloads, at list prices in US East.
1M text chunks with 768-dim embeddings, 7 GB of JSON. 1M queries a month spread through the day. 1 GB written a month.
| WarpSearch | Elastic Serverless | |
|---|---|---|
| Storage | $0.56 (7 GB held x $0.08) | $0.47 |
| Writes | $1.00 (1 GB x $1.00) | Small (ingest scales to zero) |
| Search | $2.00 (1M requests) | About $100 to $210 |
| Data returned | $0 (about 40 GB, inside the free 100 GB) | $0 (50 GB free) |
| Minimum | $10 minimum applies (usage is $3.56) | None |
| Total a month | $10 | About $100 to $210 |
Estimate from Elastic's own examples: 2 GB for $24 to $27, 20 GB for $190 to $210 a month.
5M products with 384-dim embeddings, 10 GB of JSON. 50M queries a month (about 19 a second on average). 20 GB written a month.
| WarpSearch | Elastic Serverless | |
|---|---|---|
| Storage | $0.60 (7.5 GB held x $0.08) | $0.70 |
| Writes | $20 (20 GB x $1.00) | Small |
| Search | $100 (50M requests) | $200+ (VCUs per query not published) |
| Data returned | $18 (250 GB returned, 100 GB free, before compression) | $10 |
| Minimum | Cleared | None |
| Total a month | $139 | $210+ |
Estimate. Elastic does not publish how many VCUs a query uses.
10 TB of raw logs kept searchable. 1 TB written a month. 10,000 dashboard aggregation queries a month.
| WarpSearch | Elastic Serverless | |
|---|---|---|
| Storage | $455 (8.5 TB held: 1 TB x $0.08, 7.5 TB x $0.05) | $170 to $360 |
| Writes | $1,000 (1 TB x $1.00) | $70 to $350 |
| Search | $10 (10,000 requests, about 10 CPU-seconds each) | Included |
| Data returned | $0 | $0 |
| Minimum | Cleared | None |
| Total a month | $1,465 | $240 to $710 |
Observability project. The low end uses Elastic's lowest published rates ($0.07 ingest, $0.017 retention); rates are volume tiered and the breakpoints are not published, so the high end assumes a higher starting tier. Elastic meters normalized volume, which can be larger than raw.
1 PB of raw logs kept searchable. 50 TB written a month. 100,000 heavy aggregation queries a month.
| WarpSearch | Elastic Serverless | |
|---|---|---|
| Storage | $31,280 (850 TB held, tiered) | $17,000 to $21,100 |
| Writes | $6,450 (1 TB x $1.00, 9 TB x $0.25, 40 TB x $0.08) | $3,500 to $4,800 |
| Search | $1,330 (100,000 requests, about 133 CPU-seconds each) | Included |
| Data returned | $0 | $0 |
| Minimum | Cleared | None |
| Total a month | $39,060 | $20,500 to $25,900 |
Observability project. Low end is Elastic's lowest published rates; high end assumes a tiered schedule that Elastic does not publish.
Performance
What Elastic Serverless publishes
| p99 latency, Rally wikipedia | 35 ms, 1,107 ops/s |
|---|---|
| p99 latency, MS MARCO passage (sparse) | 67 ms |
| p99 latency, vectors (BEIR NQ, 1536-dim) | 22 ms, 624 ops/s |
| Time to searchable | 5 seconds by default |
| Latency per Search Power setting | Not published |
Elastic publishes p99 numbers from its own Rally runs. The hardware is not stated, and the datasets differ from ours, so treat both sides as indicative. Our median keyword latency on similar-size text sets is 2 to 6 ms on 2 to 4 CPUs. Elastic has the more complete feature set.
What we measured
| Test | Result |
|---|---|
| Keyword search, 2.8M docs (LoTTE) | 2.1 ms median, 3,700 queries/s |
| Keyword search, 10M passages (MS MARCO v2) | 2.4 ms median (5.4 ms with an exact total hit count) |
| Keyword search, small BEIR sets | 1 to 6 ms median |
| Vector search, 10M x 128-dim (BigANN SIFT) | 0.964 recall@10 at 6.2 ms median, 162 queries/s |
| Vector search, 1M x 1536-dim (arXiv) | 0.961 recall@10 at 15.7 ms median |
| Filtered vector search, 10M (YFCC) | 0.959 recall@10 at 10.4 ms median |
| Log filters, 22M events (http_logs) | 1 to 2 ms median |
| Log aggregations, 22M events | 190 to 430 ms median (sort, daily histogram, avg by status) |
| Ingest, short text | 63,000 to 97,000 docs/s |
| Ingest, logs (247M events) | 76,000 to 81,000 events/s |
| Ingest, 768 to 1536-dim vectors | 2,200 to 4,000 vectors/s |
| Time to searchable | Seconds, or immediate with commit=true |
| Cold start | Not measured yet. First query on a 2.3 GB index: 17 to 19 ms vs 2.7 ms warm (page cache only) |
Measured October 2026 through our Elasticsearch-compatible API, warm cache, client in the same region, single runs. Every test ran on deliberately minimal hardware, 2 to 4 CPUs (a 4 CPU container or an AWS i8g.large with 2 vCPUs), so these are the least favorable results we can produce. More hardware only improves them. Numbers exclude internet round trips.
Bottom line
Choose WarpSearch for
- Small apps: $10 instead of about $100 to $210, because Elastic search projects keep a baseline running.
- Drop-in for Elasticsearch, OpenSearch and Solr clients on one endpoint.
- Per-request pricing you can predict, instead of VCUs per query that Elastic does not publish.
Choose Elastic Serverless for
- Logs: an Elastic Observability project costs about 16 to 48% of our price for mid-size logs, and 52 to 66% at petabyte scale, depending on volume tiers Elastic does not publish.
- Features: full aggregation set, RRF retrievers, ML inference, Kibana.
Pick WarpSearch for search apps with modest or bursty traffic, where Elastic's baseline costs more than our usage. Pick Elastic Observability for log retention at scale.
Sources
- Serverless Search pricing
- Serverless Observability pricing
- Serverless performance on AWS (Jan 2026)
- Project settings and Search Power
- WarpSearch pricing
Elastic lists rates as "as low as" without a region. Search-project costs depend on VCU use, which Elastic does not publish per query. Prices checked October 20, 2026. List prices in US dollars, before discounts. Estimates are marked. Tell us if something is out of date: contact us.