📈 Scalability glossary

1. ➡️ A/B Testing - allows deploying multiple versions of a service as independent instances, using load balancers or API gateways to route traffic between versions A and B Centralized observability tools capture real-time performance and user interaction data, which is aggregated for analysis to assess conversion rates and effectiveness. This setup supports safe rollbacks, enabling teams to revert to the stable version instantly if variant B underperforms, facilitating resilient, data-driven iteration.
2. 🔹 Agreed Service Time (AST) - time during which a particular Business Continuity is agreed to be fully available, ideally as defined in the Service Level Agreement.
3. 📐 Availability (Service) - measures how reliably a system is operational and accessible, typically represented as a percentage of uptime over a set period. High availability is critical, often defined in SLAs, where many organizations aim for “five nines” (99.999%), meaning minimal downtime annually. Fault tolerance and redundancy are essential strategies to achieve high availability, allowing systems to continue functioning even if components fail. Monitoring and scheduled maintenance also help by proactively identifying and addressing issues, reducing unplanned downtime.
4. 📐 Bandwidth - is the format and size of the messages that are exchanged between the microservices. You can optimize the size by applying compression, filtering, batching, caching techniques or changing the format from verbose JSON to binary Protobuf or Avro.
5. 🛠️ Benchmarking tools 
6. 🔹 Bottleneck - refer to limitations within a system that hinder its ability to grow and manage increased workloads effectively. Identifying scalability bottlenecks is crucial for improving system performance, as addressing them can lead to significant enhancements in throughput and response times.
7. ⚙️ Cache (in-memory)
8. ⚙️ Cache (multilevel) 
• L1 (Instruction/Data Level) - lightning-fast caches sit closest to the CPU cores, and their primary role is to store frequently accessed instructions and data. JVM: Ehcache; PHP: APCu/OPCache.
• L2 Application - Distributed cache systems like Redis or Memcached which allows to share data between many application instances.
• L3 DB - Materialized Views.
• L4 HTTP - Varnish.
9. 🔹 Capacity (System)
10. ⚙️ CloudFront (AWS) - is a CDN that delivers static and dynamic content reliably, securely, and efficiently.
11. ⚙️ Content Delivery Networks (CDN) - geographically distributed group of servers that caches content close to end users.
12. 🔹 Cost Upfront
13. 📐 Cost Efficiency
14. ➡️ Database Horizontal Partitioning
15. 📐 Database Load
16. ➡️ Database Partitioning
17. ➡️ Database Replication (master/slave)
18. 🔹 Database Retention
19. 🔹 Database (Retentional)
20. ⚙️ Database Shards
21. 📐 Database Slow Query
22. ➡️ Database Vertical Partitioning - in a table entire columns are separated out and put into new, distinct tables.
23. 🔹 Data Consistency (in case of microservices)
24. 🔹 Data Consistency (eventual)
25. 🔹 Data Consistency (strong)
26. 🔹 Data Integrity
27. 🔹 Data Quality
28. ➡️ Diagonal Scaling - hybrid of vertical and horizontal scaling.
29. ⚙️ Distributed System
30. 🔹 Downtime (DT) [duration/frequency]
31. 🔹 Expenditure
32. 📐 Fault Tolerance - ability of a system to continue operating without interruption despite encountering failures in one or more of its components. It is a measure of the system's resilience against disruptions (single server failure/data center outage). Strategies: redundancy, sharding, replication, load balancing, failure detection and recovery.
33. ➡️ Horizontal Scaling (scale out) - ability for an application to automatically scale by adding/reducing computing nodes as the workload increases/decreases. Usually requires a load balancer. Advantages: easy increasing capacity on the fly, limitless scaling, easier to run fault-tolerance, improve resilience. Disadvantages: complex architectural design, increased costs (resources in data center), complex debugging.
34. 🔹 Idle Resource
35. 🔹 Infrastructure as Code (IaC) - practice where infrastructure provisioning and management are performed through code. This approach automates the setup and scaling of cloud environments, ensuring consistency and eliminating manual errors.
36. 🛠️ k8s - abbreviation of Kubernetes.
37. ⚙️ [k8s] Cluster - is a set of nodes that run containerized applications. Containerizing applications packages an app with its dependences and some necessary services. They are more lightweight and flexible than virtual machines. In this way, Kubernetes clusters allow for applications to be more easily developed, moved and managed.
38. 🛠️ [k8s] Cluster Autoscaler (CA)
39. 🛠️ [k8s] Horizontal Pod Autoscaler (HPA)
40. ⚙️ [k8s] Node - is a physical or virtual machine that serves as the infrastructure where Pods (containers) run. Each Node provides essential resources, like CPU, memory, and storage, to support applications within the Kubernetes cluster. Nodes run key Kubernetes components such as the kubelet and kube-proxy to manage container operations and network communication. If a Node fails, Kubernetes automatically relocates its workloads to other Nodes, ensuring application stability and high availability across the cluster.
41. ⚙️ [k8s] Pod - is the smallest deployable unit and serves as a container for one or more application containers. All containers within a Pod share the same network and storage resources, allowing them to communicate seamlessly and exchange data. Pods are temporary; if a Pod fails, k8s will automatically replace it with a new one to maintain application stability. K8s can also scale an application by creating multiple replicas of a Pod, balancing the load and ensuring availability across different nodes.
42. 🛠️ [k8s] Karpenter - is an open-source, flexible, and high-performance Kubernetes cluster autoscaler, offering advanced scheduling and scaling capabilities. It can improve application availability and cost efficiency by rapidly launching right-sized compute resources in response to changing application load.
43. 🛠️ [k8s] Kubernetes - is used to manage and scale applications running in containers, which are small, isolated environments. It simplifies the reliable management of numerous apps and services, even when they are distributed across multiple servers. Kubernetes automates things like: starting new apps when needed. It's open-sorce. 
44. 🛠️ [k8s] Vertical Pod Autoscaler (VPA) - is an autoscaler that enables automatic CPU and memory request and limit adjustments based on historical resource usage measurements.
45. 🔹 Latency
46. ⚙️ Load Balancer
47. 🔹 Load Testing
48. 🛠️ OpenShift
49. ➡️ Optimization (proactive)
50. 🔹 Parallel Computing
51. 🔹 Performance
52. 🛠️ Prometheus - is an open-source tool used for monitoring and alerting. It collects metrics on system performance, such as CPU usage, memory, and network traffic, and stores them in a time-series database for real-time analysis. It features the PromQL query language, which enables custom metric queries and is often paired with Grafana for advanced data visualization, making it ideal for monitoring Kubernetes and cloud-based environments.
53. 🔹 Quality of Service (QoS)
54. ⚙️ Rate Limiter
55. ➡️ Read Scaling (1)
56. 🔹 Redundancy (System)
57. 🔹 Region
58. 🔹 Reliability (System)
59. 🔹 Resilience (System)
60. ➡️ Resource Allocation Strategies
61. 📐 Response Time
62. ⚙️ Round Robin
63. 🔹 Service Discovery
64. ⚙️ Sharding - is a technique of splitting a large dataset into smaller, independent segments called shards, each stored on separate servers. This approach enhances horizontal scalability and performance by distributing the load across multiple servers, allowing each shard to handle a subset of the data. By using a shard key, data is allocated to shards based on a specific criterion, such as range, hashing, or time. Sharding, while powerful, adds complexity in data management and can make cross-shard queries and transactions more challenging to handle.
65. 🔹 Service Level Agreement (SLA)
66. 🔹 Single Point of Failure (SPOF)
67. 🔹 Temporary Unavailability (System)
68. 🛠️ Terraform - Infrastructure as Code (IaC) software tool offered by HashiCorp. It facilitates provisioning and managing your infrastructure on-prem and in the cloud. It can be easily extended with the help of its plugin-based architecture.
69. ⚙️ Time Consuming Task
70. 📐 Traffic
71. 🔹 TTL (Time to Live) - is a mechanism used in computer networking to limit the lifespan or lifetime of data in a network. It was originally designed as a way to prevent network packets from circulating indefinitely in the event of routing loops. TTL is implemented as a counter or timestamp attached to data, which is decremented as it passes through network devices.
72. 🔹 Uptime
73. 🛠️ Varnish - is an HTTP reverse proxy and caching server that accelerates web applications by storing copies of frequently accessed pages or assets in memory. Positioned between users and the backend server, Varnish can serve cached content directly, reducing server load and significantly improving response times. It’s widely used on high-traffic websites for its high performance and flexible configuration options through Varnish Configuration Language (VCL).
74. ➡️ Vertical Scaling (scale up) - consists of bolstering the computing power of a single existing server, by increasing the number of CPU cores, memory and disk storage. Virtual server hardware resources are limited to a certain level. Initial costs of Vertical Scaling are lower than Horizontal. Disadvantages: SPOF, hardware limitations, downtime on update (server is offline). 
75. 🔹 Workload
76. 🔹 Workload (heavy)
77. 🔹 Workload spike
78. 🔹 Workload distribution
79. 🔹 Workload volumes
80. ➡️ Write Scaling (1)