Banco Bradesco, the third-largest banking institution by assets in Brazil and Latin America, centralized its software delivery lifecycle on Elastic Observability and Elastic Cloud on Microsoft Azure, reaching 100% visibility across its development pipelines during a migration that spans more than 70,000 GitHub repositories.
Summary
Banco Bradesco, founded in 1943 and now the third-largest banking institution by assets in Brazil and Latin America, is migrating an application portfolio backed by more than 70,000 GitHub repositories to Microsoft Azure under a modernization program called Project LEAP. Before Elastic, its development telemetry was scattered across separate tools, forcing teams to cross-reference data by hand and making lead time and delivery risk nearly impossible to measure. The bank consolidated project management, code quality, and version control data into Elastic Observability on Elastic Cloud on Azure, using Elasticsearch as a real-time data hub. It now has 100% visibility across its primary development pipelines, measures delivery against the DORA and SPACE frameworks in real time, and retains five years of history for regression analysis.
Replatforming the third-largest bank in Brazil and Latin America
Reaching 100% visibility across its software delivery pipelines was not a nice-to-have for Banco Bradesco; it was a precondition for safely replatforming one of Latin America's largest banks. Founded in 1943, Banco Bradesco has turned a long-standing commitment to its customers into a position as one of the 50 most valuable banks in the world and the third-largest banking institution by assets in Brazil and Latin America. To sustain that standing in a digital-first market, the bank is running Project LEAP, a modernization and cloud migration effort that is moving an application portfolio backed by more than 70,000 GitHub repositories to Microsoft Azure.
A transition of that magnitude carries real operational risk. In a tightly interconnected environment, even a minor software update can have far-reaching effects on other systems. The engineering challenge was to keep a constant flow of updates tracked and tested end to end, across the whole portfolio, without disrupting the experience of the millions of customers who bank with Bradesco every day.
"Until recently, the bank lacked a central platform to correlate data. Everything was decentralized. Each tool had its own metrics and everyone looked at it in a different way. By unifying our telemetry in Elastic, we established a single source of truth that aligned our technical efforts with our business goals to accelerate our delivery."
Before: Delivery data trapped in separate tools
Before adopting Elastic, Banco Bradesco's development data was fragmented. The bank used several industry-standard tools, but the information each one produced stayed locked inside that individual platform. Teams had to manually cross-reference data from one tool to the next, which introduced significant delays whenever they tried to pinpoint the source of a bottleneck or a performance lag. Without a single view of the CI/CD development cycle, measuring a metric as basic as total lead time accurately was nearly impossible.
Log management ran on a separate third-party analytics tool, used in a decentralized way. Different teams looked at metrics through different lenses, which produced inconsistent reporting and conflicting priorities. With no way to connect the data, the bank could not reliably measure the true efficiency of its software delivery or identify which stages of the migration were introducing the most operational risk.
Unifying the ecosystem on a single data hub
To close that gap, Banco Bradesco standardized on Elastic Observability, built on the Elasticsearch engine, as its real-time data hub. Moving from a self-managed, on-premises deployment to Elastic Cloud on Azure gave the bank a scalable, managed platform able to ingest and transform data from across its entire development lifecycle.
With that foundation in place, the bank now pulls data from project management tools, code quality scanners, and version control systems into a single location. Using transforms in Elasticsearch, it processes and summarizes a high volume of delivery data into a standardized format. The result is transparency: Technical telemetry stops being an abstract concept and becomes a clear map of exactly how a project is progressing and where bottlenecks are forming. That visibility is what lets the bank hold a high migration velocity while proactively mitigating operational risks before they reach the customer.
"We chose Elasticsearch because it gives the developer more freedom. We can process and transform data in ways that more rigid tools don't allow. Regardless of the programming language or the workflow, we can condense that information into a standardized format."
Technical highlights
- Real-time data hub: Elastic Observability on Elasticsearch, running on Elastic Cloud on Microsoft Azure
- Migrated from a self-managed, on-premises deployment to a managed Elastic Cloud deployment
- Ingests from project management tools, code quality scanners, and version control systems
- Elasticsearch transforms summarize high-volume delivery data into a standardized format
- ES|QL for high-speed, concurrent querying across delivery telemetry
- Real-time dashboards built with Kibana Lens and Time Series Visual Builder (TSVB)
- Index lifecycle management (ILM) retaining five years of history
- DORA and SPACE delivery frameworks measured against unified pipeline data
- Scope: Application portfolio backed by more than 70,000 GitHub repositories
Driving quality with DORA and SPACE benchmarks
With one unified data source in place, Banco Bradesco reached 100% visibility across its primary development pipelines and used that transparency to adopt two industry-standard evaluation frameworks: DORA and SPACE. Together they let the bank measure deployment frequency and the lead time for a new feature to go from an idea to a live service.
To analyze that data efficiently, the team adopted ES|QL, Elasticsearch's piped query language, which runs complex searches and concurrent processing at high speed. The team monitors the resulting metrics in real time through custom Kibana Lens and Time Series Visual Builder (TSVB) visualizations. When the data shows a spike in technical errors, engineers can step in before the update reaches production. For a bank with millions of customers using its mobile and web applications to manage their money, deploying only high-quality code directly protects the reliability customers expect.
Empowering developers with self-service data access
One of the biggest shifts has been in developer autonomy. Previously, a developer who needed to investigate a technical issue often had to file a manual request with an administrative team to get access to specific logs. That dependency slowed problem-solving down at exactly the moment speed mattered most.
Today, the bank uses Elastic's secure access controls to give developers direct access to the data they need inside Elasticsearch. The self-service model has lifted productivity and lets teams resolve incidents far faster. The bank also uses index lifecycle management to retain its data history over a five-year window, a marked change from previous systems, where logs were sometimes deleted after only a few days. That long retention lets the bank compare historical, premigration data against the new Azure environment, confirming the transition meets service requirements and troubleshooting any migration-related regressions immediately.
Predicting bottlenecks with machine learning
Banco Bradesco has begun applying Elastic machine learning to pattern recognition across its delivery data, helping teams generate predictions and flag potential bottlenecks in the software delivery pipeline before they occur. It is an early step in a deliberate move from reactive monitoring toward a proactive, centralized data strategy, one that turns the complexity of a vast ecosystem into a predictable, data-driven engine.
Before and after
| Area | Before | After |
|---|---|---|
| Delivery visibility | Data fragmented across separate tools, cross-referenced by hand | 100% visibility across primary pipelines in one platform |
| Delivery metrics | Lead time and delivery risk nearly impossible to measure | DORA and SPACE benchmarks measured in real time |
| Reporting | Each team viewed metrics through a different lens | A single source of truth aligned to business goals |
| Log access | Developers filed manual requests to administrative teams | Secure self-service access to logs and insights |
| Data retention | Logs sometimes deleted after a few days | Five years of history under index lifecycle management |
| Issue handling | Problems surfaced after the fact | Issues caught early, with ML-assisted prediction |
What comes next?
With 100% pipeline visibility and a scalable foundation on Elastic Cloud, Banco Bradesco has already turned a sprawling delivery ecosystem into a predictable, data-driven engine. The next step is making that data even easier to reach. The bank is exploring how to use Elastic AI Agent to securely connect its private development data to large language models (LLMs), so any team member could ask about a project's status in natural language and get an instant answer drawn from the bank's own data.
Your organization may not be migrating 70,000 repositories today, but the same principles apply whether you are standardizing a handful of pipelines or scaling to an entire portfolio: Unify the telemetry, measure delivery against real benchmarks, and give engineers direct, secure access to the data they need.
"We are using Elasticsearch as a centralized database for all our tools and every type of data we want to use. We have a long road ahead, but using the tool as a true data hub — not just for monitoring, but for tracking the entire lifecycle of our software — will allow us to continue delivering innovative digital services to our customers at speed."
Related resources
Topics: Software delivery observability, CI/CD, DevOps, DORA metrics, SPACE framework, Elasticsearch, ES|QL, index lifecycle management, Kibana, Elastic Cloud, Microsoft Azure, cloud migration, machine learning, Financial Services, Elastic Observability
See how Elastic Observability gives you full visibility across your software delivery pipelines, or start now with a free trial.